Not in the top 1,000 by GitHub stars
modular/skills
Agent Skills for Mojo and MAX development
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
benchmark-model/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
debug-model/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
A local comparison is unavailable. You can inspect the recorded versions on GitHub.
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
eval-model/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
import-model/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
A local comparison is unavailable. You can inspect the recorded versions on GitHub.
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
migrate-max-v2-to-v3/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
A local comparison is unavailable. You can inspect the recorded versions on GitHub.
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
mojo-gpu-fundamentals/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
A local comparison is unavailable. You can inspect the recorded versions on GitHub.
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
mojo-python-interop/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
A local comparison is unavailable. You can inspect the recorded versions on GitHub.
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
mojo-syntax/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
A local comparison is unavailable. You can inspect the recorded versions on GitHub.
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
new-modular-project/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
profile-model/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
serve-model/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Not checked yet → After Network or broad shell
9837a3af93e2 → 0e44ae625e3b
Diff text is unavailable under the recorded licenses. Line counts and the GitHub link remain available.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Not checked yet
First observation → 9837a3af93e2
No earlier version was observed for comparison.
View on GitHubAdd to my skills
Verified commit 0e44ae625e3bc0d158f3243e1636f3861eedf3a4
git clone --filter=blob:none --no-checkout https://github.com/modular/skills.git .hotskills/modular-skills && git -C .hotskills/modular-skills checkout --detach 0e44ae625e3bc0d158f3243e1636f3861eedf3a4
# Copy each selected skill directory from .hotskills/modular-skills into .claude/skills/<skill-name>/208GitHub stars
1star in 7 days
11skills in this repository
40commits in 30 days
Heat
3.1- Star growth, last 7 daysNot available
- Star growth, last 24 hoursNot available
- Rank movement, last 7 daysNot available
- Commits, last 30 daysNot available
- Accounts asking about it through MCP, last 7 daysNot available
- Hot votes this weekNot available
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
30-day change not yet available: not enough history.
Individual skills in this repository
- benchmark-model
benchmark-model - debug-model
debug-model - eval-model
eval-model - import-model
import-model - migrate-max-v2-to-v3
migrate-max-v2-to-v3 - mojo-gpu-fundamentals
mojo-gpu-fundamentals - mojo-python-interop
mojo-python-interop - mojo-syntax
mojo-syntax - new-modular-project
new-modular-project - profile-model
profile-model - serve-model
serve-model
How it works
benchmark-model
When your agent uses it
Benchmark a model served on MAX with the `max benchmark` command: measure throughput (tokens/sec), latency (TTFT, TPOT, inter-token latency), and GPU utilization by driving load against a running `max serve` endpoint. Use this whenever the user wants to benchmark, load-test, or measure the performance of a MAX model, get tokens-per-second / TTFT / TPOT numbers, run a concurrency or request-rate sweep, compare latency vs throughput, size a deployment, or produce benchmark JSON, even if they don't say "benchmark" by name. Also use when a `max benchmark` run fails to connect or reports zero/garbage numbers.
What it loads into context
Frontmatter
222 tokensAlways available to the agent
SKILL.md
2,049 tokensWhen the skill is activated
Reference documents
113,008 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- References
- 1. Check the server and read its model name
- 2. Pick the workload for your question
- 3. Run it, save results, add GPU stats
- 4. Read the metrics
- Troubleshooting
Scripts executed separately
../eval-model/scripts/_eval_tasks.py../eval-model/scripts/eval_dataset.py../eval-model/tests/test_eval_dataset.py../import-model/scripts/__init__.py../import-model/scripts/check_port.py../import-model/scripts/check_walls.py../import-model/scripts/checkpoint_metadata.py../import-model/scripts/compare_layers.py../import-model/scripts/conftest.py../import-model/scripts/donor.py../import-model/scripts/dtype_utils.py../import-model/scripts/full_copy.py../import-model/scripts/hub_config.py../import-model/scripts/import_model.py../import-model/scripts/inspect_hf.py../import-model/scripts/list_checkpoint_keys.py../import-model/scripts/list_native_archs.py../import-model/scripts/max_arch_paths.py../import-model/scripts/run_oss_gates.py../import-model/scripts/scaffold.py../import-model/scripts/templates.py../import-model/scripts/test_scripts.py../profile-model/scripts/gpu_snapshot.py../serve-model/scripts/suggest_serve_command.py
Files
../.claude-plugin/marketplace.jsonasset1,474 bytesSKILL.mdcore8,210 bytesSKILL.mdcore8,210 bytesSKILL.mdcore8,210 bytesSKILL.mdcore8,210 bytesreferences/flags.mddoc6,532 bytesreferences/metrics.mddoc4,524 bytesreferences/troubleshooting.mddoc2,506 bytes../debug-model/SKILL.mddoc7,312 bytes../debug-model/references/agent-workflow.mddoc2,670 bytes../debug-model/references/comparator-build.mddoc12,648 bytes../debug-model/references/comparator-output-patterns.mddoc4,074 bytes../debug-model/references/stacked-failures.mddoc2,881 bytes../eval-model/.gitattributesasset128 bytes../eval-model/.gitignoreasset116 bytes../eval-model/SKILL.mddoc14,110 bytes../eval-model/pixi.tomlasset449 bytes../eval-model/pixi.tomlasset449 bytes../eval-model/pixi.tomlasset449 bytes../eval-model/pixi.tomlasset449 bytes../eval-model/references/examples.mddoc3,956 bytes../eval-model/references/fast-fail-thresholds.mddoc1,698 bytes../eval-model/references/heartbeat-protocol.mddoc3,968 bytes../eval-model/scripts/_eval_tasks.pyruns25,391 bytes../eval-model/scripts/eval_dataset.pyruns35,684 bytes../eval-model/tests/test_eval_dataset.pyruns10,384 bytes../import-model/.gitattributesasset128 bytes../import-model/.gitignoreasset47 bytes../import-model/README.mddoc3,589 bytes../import-model/SKILL.mddoc30,064 bytes../import-model/pixi.tomlasset3,364 bytes../import-model/pixi.tomlasset3,364 bytes../import-model/pixi.tomlasset3,364 bytes../import-model/pixi.tomlasset3,364 bytes../import-model/pixi.tomlasset3,364 bytes../import-model/pixi.tomlasset3,364 bytes../import-model/references/distributed-transformer.mddoc10,889 bytes../import-model/references/divergences.mddoc18,044 bytes../import-model/references/implement-graph.mddoc15,698 bytes../import-model/references/layer-by-layer-debugging.mddoc7,673 bytes../import-model/references/map-to-max.mddoc5,831 bytes../import-model/references/max-vs-port-isolation.mddoc3,831 bytes../import-model/references/native-arch-check.mddoc4,410 bytes../import-model/references/pitfalls-config.mddoc6,467 bytes../import-model/references/pitfalls-graph.mddoc5,035 bytes../import-model/references/pitfalls-serving.mddoc4,516 bytes../import-model/references/pitfalls-weights.mddoc4,735 bytes../import-model/references/pitfalls.mddoc2,994 bytes../import-model/references/plan-and-veto.mddoc2,860 bytes../import-model/references/read-config-json.mddoc9,340 bytes../import-model/references/read-modeling-code.mddoc8,953 bytes../import-model/references/recognize-walls.mddoc7,698 bytes../import-model/references/rename-weights.mddoc9,901 bytes../import-model/references/rewrite-donor-docstrings.mddoc6,167 bytes../import-model/references/serve-and-iterate.mddoc5,881 bytes../import-model/references/state-dict-audit.mddoc6,186 bytes../import-model/references/template.mddoc8,327 bytes../import-model/references/tests-and-ci.mddoc2,604 bytes../import-model/references/validation-tiers.mddoc6,295 bytes../import-model/scripts/__init__.pyruns1,049 bytes../import-model/scripts/check_port.pyruns8,297 bytes../import-model/scripts/check_walls.pyruns5,317 bytes../import-model/scripts/checkpoint_metadata.pyruns4,062 bytes../import-model/scripts/compare_layers.pyruns15,852 bytes../import-model/scripts/conftest.pyruns1,650 bytes../import-model/scripts/donor.pyruns22,840 bytes../import-model/scripts/dtype_utils.pyruns2,465 bytes../import-model/scripts/full_copy.pyruns8,627 bytes../import-model/scripts/hub_config.pyruns1,589 bytes../import-model/scripts/import_model.pyruns4,204 bytes../import-model/scripts/inspect_hf.pyruns8,777 bytes../import-model/scripts/list_checkpoint_keys.pyruns3,959 bytes../import-model/scripts/list_native_archs.pyruns3,875 bytes../import-model/scripts/max_arch_paths.pyruns13,530 bytes../import-model/scripts/run_oss_gates.pyruns12,027 bytes../import-model/scripts/scaffold.pyruns12,235 bytes../import-model/scripts/templates.pyruns31,568 bytes../import-model/scripts/test_scripts.pyruns46,163 bytes../migrate-max-v2-to-v3/SKILL.mddoc11,314 bytes../migrate-max-v2-to-v3/references/module-registration.mddoc11,794 bytes../migrate-max-v2-to-v3/references/v2-v3-basics.mddoc9,013 bytes../mojo-gpu-fundamentals/SKILL.mddoc21,865 bytes../mojo-python-interop/SKILL.mddoc9,731 bytes../mojo-syntax/SKILL.mddoc30,185 bytes../mojo-syntax/references/idiomatic-mojo.mddoc29,249 bytes../new-modular-project/SKILL.mddoc6,637 bytes../profile-model/SKILL.mddoc10,554 bytes../profile-model/references/custom-architectures.mddoc5,346 bytes../profile-model/references/ncu.mddoc3,146 bytes../profile-model/references/nsys.mddoc7,214 bytes../profile-model/references/rocprofv3.mddoc3,679 bytes../profile-model/references/utilization-api.mddoc4,322 bytes../profile-model/scripts/gpu_snapshot.pyruns5,159 bytes../serve-model/SKILL.mddoc16,806 bytes../serve-model/references/custom-arch.mddoc5,987 bytes../serve-model/references/flags.mddoc7,703 bytes../serve-model/references/troubleshooting.mddoc4,764 bytes../serve-model/scripts/suggest_serve_command.pyruns11,945 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubdebug-model
When your agent uses it
Debug a MAX model that loads, compiles, serves, and generates tokens but whose output disagrees with a reference implementation. Use whenever parity debugging stalls on scalar taps, the model returns gibberish or wrong greedy tokens, logit cosine is high but argmax differs, or generation is coherent then diverges. Applies during an architecture port, a quantization bring-up, a multi-GPU conversion, or after a MAX upgrade. Triggers on "parity failure", "silent corruption", "logits match but tokens diverge", "top-1 mismatch", "greedy divergence", and "model serves but generates garbage". Not for crashes on load or pre-serve scaffolding (use import-model). Mandates reference-vs-MAX tensor-dump comparators first, numerical verification of fixes before recompiling, and a serve-vs-pipeline bisect when dumps match but text diverges.
What it loads into context
Frontmatter
238 tokensAlways available to the agent
SKILL.md
1,625 tokensWhen the skill is activated
Reference documents
5,377 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- References
- Protocol
Files
SKILL.mdcore7,312 bytesSKILL.mdcore7,312 bytesSKILL.mdcore7,312 bytesSKILL.mdcore7,312 bytesreferences/agent-workflow.mddoc2,670 bytesreferences/comparator-build.mddoc12,648 bytesreferences/comparator-output-patterns.mddoc4,074 bytesreferences/stacked-failures.mddoc2,881 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubeval-model
When your agent uses it
Measures the task accuracy of text models served by MAX using standard benchmarks such as GSM8K, MMLU, HellaSwag, ARC, AIME, GPQA, TruthfulQA, WinoGrande, and BABILong. Use when benchmarking a served model, comparing it with model-card or reference scores, verifying that a new MAX model produces correct answers, or running repeatable dataset evaluations against a MAX OpenAI-compatible endpoint.
What it loads into context
Frontmatter
153 tokensAlways available to the agent
SKILL.md
3,510 tokensWhen the skill is activated
Reference documents
2,689 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- References
- Set up the evaluator
- Serve the model
- Run an evaluation
- Interpret the result
- Choose tasks
- Evaluate reasoning models
- Monitor progress
- Troubleshoot failures
Scripts executed separately
scripts/_eval_tasks.pyscripts/eval_dataset.pytests/test_eval_dataset.py
Files
.gitattributesasset128 bytes.gitignoreasset116 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytesSKILL.mdcore14,110 bytespixi.tomlasset449 bytesreferences/examples.mddoc3,956 bytesreferences/fast-fail-thresholds.mddoc1,698 bytesreferences/heartbeat-protocol.mddoc3,968 bytesscripts/_eval_tasks.pyruns25,391 bytesscripts/eval_dataset.pyruns35,684 bytestests/test_eval_dataset.pyruns10,384 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubimport-model
When your agent uses it
Use when importing a new model architecture into MAX from a Hugging Face model ID. Triggers on: "import a model into MAX", "add model to MAX", "bring up <HF model> in MAX". Workflow: inspect the Hugging Face config and modeling code, then scaffold from a similar ModuleV3 architecture. Implement each layer as a max.nn Module that matches HF, serve, then verify against the Hugging Face reference. When the server runs but output is wrong (gibberish, greedy mismatch, coherent-then-diverges), load debug-model for the divergence hunt.
What it loads into context
Frontmatter
192 tokensAlways available to the agent
SKILL.md
7,266 tokensWhen the skill is activated
Reference documents
43,605 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Phase 1: Decide and plan
- Phase 2: Implement
- Phase 3: Verify
- Common pitfalls
- Tests and CI
Scripts executed separately
scripts/__init__.pyscripts/check_port.pyscripts/check_walls.pyscripts/checkpoint_metadata.pyscripts/compare_layers.pyscripts/conftest.pyscripts/donor.pyscripts/dtype_utils.pyscripts/full_copy.pyscripts/hub_config.pyscripts/import_model.pyscripts/inspect_hf.pyscripts/list_checkpoint_keys.pyscripts/list_native_archs.pyscripts/max_arch_paths.pyscripts/run_oss_gates.pyscripts/scaffold.pyscripts/templates.pyscripts/test_scripts.py
Files
../debug-model/SKILL.mddoc7,312 bytes.gitattributesasset128 bytes.gitignoreasset47 bytesREADME.mddoc3,589 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytesSKILL.mdcore30,064 bytespixi.tomlasset3,364 bytesreferences/distributed-transformer.mddoc10,889 bytesreferences/divergences.mddoc18,044 bytesreferences/implement-graph.mddoc15,698 bytesreferences/layer-by-layer-debugging.mddoc7,673 bytesreferences/map-to-max.mddoc5,831 bytesreferences/max-vs-port-isolation.mddoc3,831 bytesreferences/native-arch-check.mddoc4,410 bytesreferences/pitfalls-config.mddoc6,467 bytesreferences/pitfalls-graph.mddoc5,035 bytesreferences/pitfalls-serving.mddoc4,516 bytesreferences/pitfalls-weights.mddoc4,735 bytesreferences/pitfalls.mddoc2,994 bytesreferences/plan-and-veto.mddoc2,860 bytesreferences/read-config-json.mddoc9,340 bytesreferences/read-modeling-code.mddoc8,953 bytesreferences/recognize-walls.mddoc7,698 bytesreferences/rename-weights.mddoc9,901 bytesreferences/rewrite-donor-docstrings.mddoc6,167 bytesreferences/serve-and-iterate.mddoc5,881 bytesreferences/state-dict-audit.mddoc6,186 bytesreferences/template.mddoc8,327 bytesreferences/tests-and-ci.mddoc2,604 bytesreferences/validation-tiers.mddoc6,295 bytesscripts/__init__.pyruns1,049 bytesscripts/check_port.pyruns8,297 bytesscripts/check_walls.pyruns5,317 bytesscripts/checkpoint_metadata.pyruns4,062 bytesscripts/compare_layers.pyruns15,852 bytesscripts/conftest.pyruns1,650 bytesscripts/donor.pyruns22,840 bytesscripts/dtype_utils.pyruns2,465 bytesscripts/full_copy.pyruns8,627 bytesscripts/hub_config.pyruns1,589 bytesscripts/import_model.pyruns4,204 bytesscripts/inspect_hf.pyruns8,777 bytesscripts/list_checkpoint_keys.pyruns3,959 bytesscripts/list_native_archs.pyruns3,875 bytesscripts/max_arch_paths.pyruns13,530 bytesscripts/run_oss_gates.pyruns12,027 bytesscripts/scaffold.pyruns12,235 bytesscripts/templates.pyruns31,568 bytesscripts/test_scripts.pyruns46,163 bytes../migrate-max-v2-to-v3/references/module-registration.mddoc11,794 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmigrate-max-v2-to-v3
When your agent uses it
Use when migrating a MAX model from the V2 graph API (max.nn.legacy, TensorValue, explicit Graph construction) to ModuleV3 (max.nn, Tensor, F.lazy() + compile()). Triggers on: "migrate to ModuleV3", "port a max.nn.legacy model to max.experimental", "convert TensorValue to Tensor", "update my model to the V3 API", "ModuleV3 migration". Workflow: read the concept map, rewrite imports and __call__ signatures, and replace graph building with compile(weights=...). Align weight names with the new module hierarchy and move the KV cache unflatten into forward(). Port sharding to a DeviceMesh when the V2 model is sharded, then verify greedy outputs match the V2 implementation.
What it loads into context
Frontmatter
238 tokensAlways available to the agent
SKILL.md
2,839 tokensWhen the skill is activated
Reference documents
4,206 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Reference implementations
- When there is no `Tensor` equivalent
Files
SKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesSKILL.mdcore11,314 bytesreferences/module-registration.mddoc11,794 bytesreferences/v2-v3-basics.mddoc9,013 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmojo-gpu-fundamentals
When your agent uses it
The basics of how to program GPUs using Mojo. Use this skill in addition to mojo-syntax when writing Mojo code that targets GPUs or other accelerators. Use targeting code to NVIDIA, AMD, Apple silicon GPUs, or others. Use this skill to overcome misconceptions about how Mojo GPU code is written.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
5,521 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Not-CUDA — key concept mapping
- Imports
- Kernel definition
- TileTensor — the primary GPU data abstraction
- Memory management
- Kernel launch
- Shared memory
- Thread indexing
- Synchronization and warp operations
- GPU availability check
- Architecture detection — `is_` vs `has_`
- Compile-time constants pattern
- Complete 1D example (vector addition)
- Complete 2D example (tiled matmul with shared memory)
- SIMD loads in kernels
- Reduction pattern
- DeviceBuffer from existing pointer
- Benchmarking GPU kernels
- Hardware details
Files
SKILL.mdcore21,865 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmojo-python-interop
When your agent uses it
Aids in writing Mojo code that interoperates with Python using current syntax and conventions. Use this skill in addition to mojo-syntax when writing Mojo code that interacts with Python, calls Python libraries from Mojo, or exposes Mojo types/functions to Python. Also use when the user wants to build Python extension modules in Mojo, wrap Mojo structs for Python consumption, or convert between Python and Mojo types.
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
2,334 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Using Python from Mojo
- Calling Mojo from Python (extension modules)
Files
SKILL.mdcore9,731 bytesSKILL.mdcore9,731 bytesSKILL.mdcore9,731 bytesSKILL.mdcore9,731 bytesSKILL.mdcore9,731 bytesSKILL.mdcore9,731 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmojo-syntax
When your agent uses it
Help to write Mojo code using current syntax and conventions. Always use this skill when writing any Mojo code, including when other Mojo-specific skills (e.g., mojo-gpu-fundamentals) also apply. Use when writing or reviewing Mojo code, translating projects to Mojo, or otherwise generating Mojo. Use this skill to overcome misconceptions with how Mojo is written.
What it loads into context
Frontmatter
80 tokensAlways available to the agent
SKILL.md
7,127 tokensWhen the skill is activated
Reference documents
7,562 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Removed syntax — DO NOT generate these
- `var` is required for every new declaration
- `def` is the only function keyword
- Mojo functions that raise must be marked as such
- `comptime` replaces `alias` and `@parameter if`/`for`
- Argument conventions
- Lifecycle methods
- Struct patterns
- Imports use `std.` prefix
- `Writable` / `Writer` (replaces `Stringable`)
- Iterator protocol
- Memory and pointer types
- Origin system (not "lifetime")
- Testing
- Dict iteration
- Collection literals
- Variant access
- Common decorators
- Contextual member references — prefer `.member`
- Numeric conversions — must be explicit
- SIMD operations
- Strings
- Error handling
- Function types and closures
- Type hierarchy
Files
SKILL.mdcore30,185 bytesSKILL.mdcore30,185 bytesSKILL.mdcore30,185 bytesreferences/idiomatic-mojo.mddoc29,249 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubnew-modular-project
When your agent uses it
Creates a new Mojo or MAX project. Use when wanting to start a new Mojo or MAX project, initializing the `pixi` or `uv` environment to use Mojo or MAX, or when the user wants to begin a new Mojo or MAX project from scratch.
What it loads into context
Frontmatter
62 tokensAlways available to the agent
SKILL.md
1,729 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- System prerequisites
- Pixi (recommended)
- uv
- pip
- Conda
- Version alignment with MAX
- References
Files
SKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytesSKILL.mdcore6,637 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubprofile-model
When your agent uses it
Profile a model running on MAX to find where it spends time and whether the GPU is saturated. Use when the user asks to "profile my model," "where is my model spending time," "why is inference slow," "is my GPU being utilized," "how much GPU am I using," "get a kernel breakdown," "capture an nsys/rocprof/ncu trace of max serve," or wants to measure MAX inference performance. Works for any model MAX can run — built-in architectures and custom ones loaded with --custom-architectures — from a pip or pixi install (max generate, max serve, or a Python script) on NVIDIA or AMD GPUs. Decide cheapest-first: a GPU utilization check, then a kernel breakdown, then a single-kernel deep dive only when one kernel dominates.
What it loads into context
Frontmatter
250 tokensAlways available to the agent
SKILL.md
2,636 tokensWhen the skill is activated
Reference documents
6,290 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- How to decide what to run
- Reading the result (what the kernel mix tells you)
- The durable algorithm
- Nightly vs stable
- Install notes
- Fast paths
Scripts executed separately
scripts/gpu_snapshot.py
Files
SKILL.mdcore10,554 bytesSKILL.mdcore10,554 bytesreferences/custom-architectures.mddoc5,346 bytesreferences/ncu.mddoc3,146 bytesreferences/nsys.mddoc7,214 bytesreferences/rocprofv3.mddoc3,679 bytesreferences/utilization-api.mddoc4,322 bytesscripts/gpu_snapshot.pyruns5,159 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubserve-model
When your agent uses it
Serve a model with MAX's `max serve` command: set up the environment (pixi or uv with the max-nightly conda channel / nightly wheel index), point the server at a Hugging Face repo or local checkpoint, target a custom architecture with `--custom-architectures`, and pick the right serve flags for the model. Use this whenever the user wants to run, launch, start, or host a model on MAX, bring up an OpenAI-compatible endpoint, serve a custom/ported architecture, debug a `max serve` startup failure, or figure out which serve flags (devices, quantization-encoding, max-length, task, trust-remote-code) a given model needs, even if they don't say "max serve" by name.
What it loads into context
Frontmatter
213 tokensAlways available to the agent
SKILL.md
4,118 tokensWhen the skill is activated
Reference documents
4,737 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- References
- Fast path (custom architecture): do this first
- 1. Make sure MAX is installed
- 2. Is this a built-in model or a custom architecture?
- 3. Target a custom architecture
- 4. Build the serve command (default-first)
- 5. Launch and confirm it works
- Cache weights and compilation for faster re-serves
- Troubleshooting
Scripts executed separately
scripts/suggest_serve_command.py
Files
SKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesSKILL.mdcore16,806 bytesreferences/custom-arch.mddoc5,987 bytesreferences/flags.mddoc7,703 bytesreferences/troubleshooting.mddoc4,764 bytesscripts/suggest_serve_command.pyruns11,945 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubSkills included
- benchmark-model
- debug-model
- eval-model
- import-model
- migrate-max-v2-to-v3
- mojo-gpu-fundamentals
- mojo-python-interop
- mojo-syntax
- new-modular-project
- profile-model
- serve-model
Repository facts
- License
- No license detected
- Created
- Feb 25, 2026
- Last pushed
- Nov 5, 2026
- Forks
- 23
- Open issues
- 4
- Project activity
- Active
Before you install
Network or broad shellAutomated signals, not a guarantee
Checked Nov 5, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 93 notes
benchmark-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedbenchmark-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedbenchmark-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzedimport-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzedimport-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzedimport-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzedimport-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzedimport-model/pixi.toml:1Analysis incomplete: Referenced executable target unavailable; not analyzeddebug-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddebug-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddebug-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedeval-model/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- No broad tool permissions
- 12 dangerous shell commandsin 4 files: import-model/scripts/test_scripts.py:742 new-modular-project/SKILL.md:44 profile-model/SKILL.md:171 and 1 more
- 457 network referencesin 45 files: new-modular-project/SKILL.md:49 migrate-max-v2-to-v3/references/module-registration.md:131 serve-model/SKILL.md:71 and 42 more
- No hidden or obfuscated text
- No instruction overrides
- 48 executable scriptsin 24 files: eval-model/scripts/_eval_tasks.py:1 eval-model/scripts/eval_dataset.py:1 eval-model/tests/test_eval_dataset.py:1 and 21 more
- 28 references to secretsin 6 files: debug-model/references/comparator-build.md:27 eval-model/scripts/_eval_tasks.py:331 eval-model/scripts/eval_dataset.py:170 and 3 more
- No requests to vote in Hot
Show all 545 signals by file
benchmark-model/SKILL.md
- Network accessline 58
curl -s http://localhost:8000/v1/health # 200 = ready - Network accessline 59
curl -s http://localhost:8000/v1/models # note the served model name - Network accessline 120
pixi run max benchmark --backend modular --base-url http://localhost:8000 \ - Network accessline 131
pixi run max benchmark --backend modular --base-url http://localhost:8000 \ - Network accessline 168
connection failures, model-name mismatches, tokenizer-alias errors, chat-template 400s, flat throughput from a batch-size cap, and warmup-skewed first-token times. Confirm that `curl /v1/health` returns 200 before you check - Network accessline 169
chat-template 400s, flat throughput from a batch-size cap, and warmup-skewed first-token times. Confirm that `curl /v1/health` returns 200 before you check anything else. - Network accessline 170
first-token times. Confirm that `curl /v1/health` returns 200 before you check
benchmark-model/references/flags.md
- Network accessline 22
`max serve` endpoint. - `--model`: a Hugging Face ID or local path. This **must equal** the server's `--served-model-name` or the requests fail. Read it from `curl /v1/models`. - Network accessline 23
- `--model`: a Hugging Face ID or local path. This **must equal** the server's `--served-model-name` or the requests fail. Read it from `curl /v1/models`. - `--endpoint`: either `/v1/completions` or `/v1/chat/completions` (default - Network accessline 24
`--served-model-name` or the requests fail. Read it from `curl /v1/models`.
benchmark-model/references/troubleshooting.md
- Network accessline 4
failures come down to one of two things: the client can't reach the server, or the workload doesn't match how you started the server. Confirm that `curl http://localhost:8000/v1/health` returns 200 before you chase anything - Network accessline 5
the workload doesn't match how you started the server. Confirm that `curl http://localhost:8000/v1/health` returns 200 before you chase anything else, because a down or still-compiling server explains most zero and garbage - Network accessline 6
`curl http://localhost:8000/v1/health` returns 200 before you chase anything - Network accessline 12
| Symptom | Cause | Fix | |---|---|---| | `Connection refused` or can't reach server | Nothing serving, or wrong host or port | Confirm `curl /v1/health` returns 200; check `--host`, `--port`, and `--base-url` against the running server | - Network accessline 13
|---|---|---| | `Connection refused` or can't reach server | Nothing serving, or wrong host or port | Confirm `curl /v1/health` returns 200; check `--host`, `--port`, and `--base-url` against the running server | | All requests fail, 404, o - Network accessline 14
| `Connection refused` or can't reach server | Nothing serving, or wrong host or port | Confirm `curl /v1/health` returns 200; check `--host`, `--port`, and `--base-url` against the running server | - Network accessline 15
| All requests fail, 404, or model not found | `--model` doesn't match the server | Read the served name from `curl /v1/models` and pass it as `--model`; it must equal `--served-model-name` |
debug-model/SKILL.md
- Network accessline 46
For MAX's built-in runtime debugging options (NaN checks, source tracebacks, op logging), see [the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 46
For MAX's built-in runtime debugging options (NaN checks, source tracebacks, op logging), see [the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 46
For MAX's built-in runtime debugging options (NaN checks, source tracebacks, op logging), see [the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 47
op logging), see [the MAX debugging tools](https://max.modular.com/develop/debugging/). `max.nn.legacy.hooks.PrintHook` (covered in comparator-build.md) prints layer - Network accessline 47
op logging), see [the MAX debugging tools](https://max.modular.com/develop/debugging/). `max.nn.legacy.hooks.PrintHook` (covered in comparator-build.md) prints layer - Network accessline 47
op logging), see [the MAX debugging tools](https://max.modular.com/develop/debugging/). `max.nn.legacy.hooks.PrintHook` (covered in comparator-build.md) prints layer - Network accessline 48
[the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 48
[the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 48
[the MAX debugging tools](https://max.modular.com/develop/debugging/).
debug-model/references/comparator-build.md
- Access to secretsline 27
[Credential-related source omitted] - Access to secretsline 27
[Credential-related source omitted] - Access to secretsline 93
[Credential-related source omitted] - Access to secretsline 93
[Credential-related source omitted] - Access to secretsline 176
[Credential-related source omitted] - Access to secretsline 176
[Credential-related source omitted] - Network accessline 218
comparison across a full run, use the saved-dump comparator above. For MAX's other built-in debugging options, see [the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 218
comparison across a full run, use the saved-dump comparator above. For MAX's other built-in debugging options, see [the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 220
[the MAX debugging tools](https://max.modular.com/develop/debugging/). - Network accessline 220
[the MAX debugging tools](https://max.modular.com/develop/debugging/). - Access to secretsline 227
[Credential-related source omitted] - Access to secretsline 227
[Credential-related source omitted] - Access to secretsline 228
[Credential-related source omitted] - Access to secretsline 228
[Credential-related source omitted]
eval-model/SKILL.md
eval-model/references/examples.md
- Network accessline 135
[Credential-related source omitted] - Network accessline 135
[Credential-related source omitted] - Access to secretsline 135
[Credential-related source omitted] - Access to secretsline 135
[Credential-related source omitted] - Network accessline 136
pixi run --manifest-path "$SKILL_DIR/pixi.toml" eval \ --base-url "https://example.com/v1" \ --model "<MODEL_ID>" \ - Network accessline 136
pixi run --manifest-path "$SKILL_DIR/pixi.toml" eval \ --base-url "https://example.com/v1" \ --model "<MODEL_ID>" \ - Network accessline 137
--base-url "https://example.com/v1" \ - Network accessline 137
--base-url "https://example.com/v1" \
eval-model/scripts/_eval_tasks.py
- Executable scriptline 1
#!/usr/bin/env python3 - Executable scriptline 1
#!/usr/bin/env python3 - Network accessline 4
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 5
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 6
# https://llvm.org/LICENSE.txt - Network accessline 6
# https://llvm.org/LICENSE.txt - Access to secretsline 331
[Credential-related source omitted] - Access to secretsline 331
[Credential-related source omitted] - Network accessline 332
[Credential-related source omitted] - Network accessline 332
[Credential-related source omitted] - Access to secretsline 332
[Credential-related source omitted] - Access to secretsline 332
[Credential-related source omitted] - Network accessline 333
[Credential-related source omitted] - Network accessline 333
[Credential-related source omitted] - Access to secretsline 333
[Credential-related source omitted] - Access to secretsline 333
[Credential-related source omitted] - Network accessline 334
req = urllib.request.Request(url, data=data, headers=headers) - Network accessline 334
req = urllib.request.Request(url, data=data, headers=headers) - Network accessline 335
try: with urllib.request.urlopen(req, timeout=timeout) as resp: return json.load(resp) - Network accessline 335
try: with urllib.request.urlopen(req, timeout=timeout) as resp: return json.load(resp) - Network accessline 336
with urllib.request.urlopen(req, timeout=timeout) as resp: - Network accessline 336
with urllib.request.urlopen(req, timeout=timeout) as resp:
eval-model/scripts/eval_dataset.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt - Access to secretsline 170
[Credential-related source omitted] - Access to secretsline 170
[Credential-related source omitted] - Network accessline 171
[Credential-related source omitted] - Network accessline 171
[Credential-related source omitted] - Access to secretsline 171
[Credential-related source omitted] - Access to secretsline 171
[Credential-related source omitted] - Network accessline 172
[Credential-related source omitted] - Network accessline 172
[Credential-related source omitted] - Access to secretsline 172
[Credential-related source omitted] - Access to secretsline 172
[Credential-related source omitted] - Network accessline 173
request = urllib.request.Request(url, data=data, headers=headers) - Network accessline 173
request = urllib.request.Request(url, data=data, headers=headers) - Network accessline 174
try: with urllib.request.urlopen(request, timeout=timeout) as response: result = json.load(response) - Network accessline 174
try: with urllib.request.urlopen(request, timeout=timeout) as response: result = json.load(response) - Network accessline 175
with urllib.request.urlopen(request, timeout=timeout) as response: - Network accessline 175
with urllib.request.urlopen(request, timeout=timeout) as response: - Network accessline 963
normalize_base_url(args.base_url) if args.base_url else f"http://{args.host}:{args.port}" - Network accessline 963
normalize_base_url(args.base_url) if args.base_url else f"http://{args.host}:{args.port}" - Network accessline 964
if args.base_url else f"http://{args.host}:{args.port}" ) - Network accessline 964
if args.base_url else f"http://{args.host}:{args.port}" ) - Network accessline 965
else f"http://{args.host}:{args.port}" - Network accessline 965
else f"http://{args.host}:{args.port}"
eval-model/tests/test_eval_dataset.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 95
) cls.thread.start() cls.base_url = f"http://127.0.0.1:{cls.server.server_port}" - Network accessline 95
) cls.thread.start() cls.base_url = f"http://127.0.0.1:{cls.server.server_port}" - Network accessline 97
cls.base_url = f"http://127.0.0.1:{cls.server.server_port}" - Network accessline 97
cls.base_url = f"http://127.0.0.1:{cls.server.server_port}" - Network accessline 105
def test_normalize_base_url_accepts_v1(self) -> None: self.assertEqual( eval_dataset.normalize_base_url("http://localhost:8000/v1/"), - Network accessline 105
def test_normalize_base_url_accepts_v1(self) -> None: self.assertEqual( eval_dataset.normalize_base_url("http://localhost:8000/v1/"), - Network accessline 106
self.assertEqual( eval_dataset.normalize_base_url("http://localhost:8000/v1/"), "http://localhost:8000", - Network accessline 106
self.assertEqual( eval_dataset.normalize_base_url("http://localhost:8000/v1/"), "http://localhost:8000", - Network accessline 107
eval_dataset.normalize_base_url("http://localhost:8000/v1/"), - Network accessline 107
eval_dataset.normalize_base_url("http://localhost:8000/v1/"), - Network accessline 108
"http://localhost:8000", - Network accessline 108
"http://localhost:8000",
import-model/SKILL.md
- Network accessline 34
Before choosing a donor, read the [module registration compatibility reference](../migrate-max-v2-to-v3/references/module-registration.md) ([published copy](https://github.com/modular/skills/blob/main/migrate-max-v2-to-v3/references/module- - Network accessline 34
Before choosing a donor, read the [module registration compatibility reference](../migrate-max-v2-to-v3/references/module-registration.md) ([published copy](https://github.com/modular/skills/blob/main/migrate-max-v2-to-v3/references/module- - Network accessline 35
[module registration compatibility reference](../migrate-max-v2-to-v3/references/module-registration.md) ([published copy](https://github.com/modular/skills/blob/main/migrate-max-v2-to-v3/references/module-registration.md) if this skill was - Network accessline 35
[module registration compatibility reference](../migrate-max-v2-to-v3/references/module-registration.md) ([published copy](https://github.com/modular/skills/blob/main/migrate-max-v2-to-v3/references/module-registration.md) if this skill was - Network accessline 36
([published copy](https://github.com/modular/skills/blob/main/migrate-max-v2-to-v3/references/module-registration.md) - Network accessline 36
([published copy](https://github.com/modular/skills/blob/main/migrate-max-v2-to-v3/references/module-registration.md) - Network accessline 166
Open `https://huggingface.co/<HF_MODEL_ID>` and read the model card for: - Network accessline 166
Open `https://huggingface.co/<HF_MODEL_ID>` and read the model card for:
import-model/pixi.toml
- Network accessline 2
name = "import-model" version = "0.1.0" channels = ["https://conda.modular.com/max-nightly", "conda-forge"] - Network accessline 2
name = "import-model" version = "0.1.0" channels = ["https://conda.modular.com/max-nightly", "conda-forge"] - Network accessline 3
version = "0.1.0" channels = ["https://conda.modular.com/max-nightly", "conda-forge"] # MAX runs on CPU, NVIDIA and AMD GPUs, and Apple silicon GPUs without a CUDA - Network accessline 3
version = "0.1.0" channels = ["https://conda.modular.com/max-nightly", "conda-forge"] # MAX runs on CPU, NVIDIA and AMD GPUs, and Apple silicon GPUs without a CUDA - Network accessline 4
channels = ["https://conda.modular.com/max-nightly", "conda-forge"] - Network accessline 4
channels = ["https://conda.modular.com/max-nightly", "conda-forge"]
import-model/references/native-arch-check.md
- Access to secretsline 9
[Credential-related source omitted] - Access to secretsline 9
[Credential-related source omitted] - Network accessline 58
Or open the raw Hub file: `https://huggingface.co/<HF_MODEL_ID>/raw/main/config.json` and read the - Network accessline 58
Or open the raw Hub file: `https://huggingface.co/<HF_MODEL_ID>/raw/main/config.json` and read the - Network accessline 59
`https://huggingface.co/<HF_MODEL_ID>/raw/main/config.json` and read the - Network accessline 59
`https://huggingface.co/<HF_MODEL_ID>/raw/main/config.json` and read the
import-model/references/pitfalls-serving.md
- Network accessline 28
for n in 1 16 64 200 2000; do curl -s http://localhost:8000/v1/completions ... -d "{\"max_tokens\": $n, ...}" - Network accessline 28
for n in 1 16 64 200 2000; do curl -s http://localhost:8000/v1/completions ... -d "{\"max_tokens\": $n, ...}" - Network accessline 29
curl -s http://localhost:8000/v1/completions ... -d "{\"max_tokens\": $n, ...}" - Network accessline 29
curl -s http://localhost:8000/v1/completions ... -d "{\"max_tokens\": $n, ...}"
import-model/references/serve-and-iterate.md
import-model/references/template.md
import-model/references/validation-tiers.md
- Network accessline 16
# MAX: served at the released weight encoding curl -s http://localhost:8000/v1/completions -H 'Content-Type: application/json' \ - Network accessline 16
# MAX: served at the released weight encoding curl -s http://localhost:8000/v1/completions -H 'Content-Type: application/json' \ - Network accessline 17
curl -s http://localhost:8000/v1/completions -H 'Content-Type: application/json' \ - Network accessline 17
curl -s http://localhost:8000/v1/completions -H 'Content-Type: application/json' \ - Network accessline 59
| Tier | What it tests | Tool | Pass criterion - Network accessline 59
| Tier | What it tests | Tool | Pass criterion - Network accessline 60
|---------------------------|-------------------------------------------------------------------------------------------|-------------------------------------------------|--------------------------------------------------------------------- - Network accessline 60
|---------------------------|-------------------------------------------------------------------------------------------|-------------------------------------------------|--------------------------------------------------------------------- - Network accessline 61
| 1: Smoke | Greedy text on one prompt looks plausible | ``curl /v1/completions`` | Real tokens, not garbage - Network accessline 61
| 1: Smoke | Greedy text on one prompt looks plausible | ``curl /v1/completions`` | Real tokens, not garbage
import-model/scripts/__init__.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/check_port.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/check_walls.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/checkpoint_metadata.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/compare_layers.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 200
} ).encode() req = urllib.request.Request( - Network accessline 200
} ).encode() req = urllib.request.Request( - Network accessline 201
).encode() req = urllib.request.Request( f"http://localhost:{port}/v1/completions", - Network accessline 201
).encode() req = urllib.request.Request( f"http://localhost:{port}/v1/completions", - Network accessline 202
req = urllib.request.Request( - Network accessline 202
req = urllib.request.Request( - Network accessline 203
f"http://localhost:{port}/v1/completions", - Network accessline 203
f"http://localhost:{port}/v1/completions", - Network accessline 205
headers={"Content-Type": "application/json"}, ) with urllib.request.urlopen(req, timeout=120) as r: - Network accessline 205
headers={"Content-Type": "application/json"}, ) with urllib.request.urlopen(req, timeout=120) as r: - Network accessline 206
) with urllib.request.urlopen(req, timeout=120) as r: choice = json.loads(r.read())["choices"][0] - Network accessline 206
) with urllib.request.urlopen(req, timeout=120) as r: choice = json.loads(r.read())["choices"][0] - Network accessline 207
with urllib.request.urlopen(req, timeout=120) as r: - Network accessline 207
with urllib.request.urlopen(req, timeout=120) as r:
import-model/scripts/conftest.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/donor.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/dtype_utils.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/full_copy.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/hub_config.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 35
return f"https://huggingface.co/{hf_id}/raw/{revision}/config.json" - Network accessline 35
return f"https://huggingface.co/{hf_id}/raw/{revision}/config.json"
import-model/scripts/import_model.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/inspect_hf.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/list_checkpoint_keys.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/list_native_archs.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 51
MAX is not installed in this Python environment (or no architectures could be discovered). Install MAX with pixi, not pip: https://max.modular.com/get-started - Network accessline 51
MAX is not installed in this Python environment (or no architectures could be discovered). Install MAX with pixi, not pip: https://max.modular.com/get-started - Network accessline 52
Install MAX with pixi, not pip: https://max.modular.com/get-started Then rerun inside that environment, for example: - Network accessline 52
Install MAX with pixi, not pip: https://max.modular.com/get-started Then rerun inside that environment, for example: - Network accessline 53
https://max.modular.com/get-started - Network accessline 53
https://max.modular.com/get-started
import-model/scripts/max_arch_paths.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 359
sys.exit( "MAX is not installed in this Python environment. " "Install MAX with pixi (https://max.modular.com/get-started), " - Network accessline 359
sys.exit( "MAX is not installed in this Python environment. " "Install MAX with pixi (https://max.modular.com/get-started), " - Network accessline 360
"MAX is not installed in this Python environment. " "Install MAX with pixi (https://max.modular.com/get-started), " "not pip." - Network accessline 360
"MAX is not installed in this Python environment. " "Install MAX with pixi (https://max.modular.com/get-started), " "not pip." - Network accessline 361
"Install MAX with pixi (https://max.modular.com/get-started), " - Network accessline 361
"Install MAX with pixi (https://max.modular.com/get-started), "
import-model/scripts/run_oss_gates.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/scaffold.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/templates.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
import-model/scripts/test_scripts.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 612
"a = huggingface_config.rope_theta\n" "b = self.huggingface_config.rope_theta\n" "c = config.rope_theta\n" - Network accessline 612
"a = huggingface_config.rope_theta\n" "b = self.huggingface_config.rope_theta\n" "c = config.rope_theta\n" - Network accessline 613
"b = self.huggingface_config.rope_theta\n" "c = config.rope_theta\n" "hf_config.rope_theta = 1.0\n" - Network accessline 613
"b = self.huggingface_config.rope_theta\n" "c = config.rope_theta\n" "hf_config.rope_theta = 1.0\n" - Dangerous shell operationline 742
exec(compile(source, "model_config.py", "exec"), namespace) - Dangerous shell operationline 742
exec(compile(source, "model_config.py", "exec"), namespace) - Dangerous shell operationline 930
exec(compile(source, "weight_adapters.py", "exec"), namespace) - Dangerous shell operationline 930
exec(compile(source, "weight_adapters.py", "exec"), namespace) - Network accessline 1160
def _http_error(body: bytes) -> urllib.error.HTTPError: return urllib.error.HTTPError( "http://localhost:8000/v1/completions", - Network accessline 1160
def _http_error(body: bytes) -> urllib.error.HTTPError: return urllib.error.HTTPError( "http://localhost:8000/v1/completions", - Network accessline 1161
return urllib.error.HTTPError( "http://localhost:8000/v1/completions", 400, - Network accessline 1161
return urllib.error.HTTPError( "http://localhost:8000/v1/completions", 400, - Network accessline 1162
"http://localhost:8000/v1/completions", - Network accessline 1162
"http://localhost:8000/v1/completions",
migrate-max-v2-to-v3/references/module-registration.md
- Network accessline 131
- [arch_lookup.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/arch_lookup.py): - Network accessline 131
- [arch_lookup.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/arch_lookup.py): - Network accessline 131
- [arch_lookup.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/arch_lookup.py): - Network accessline 132
`SupportedArchitecture` and generated pipeline selection. - [registration_graph_api.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_graph_api.py): task/backend dispatcher and graph-API decoder fa - Network accessline 132
`SupportedArchitecture` and generated pipeline selection. - [registration_graph_api.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_graph_api.py): task/backend dispatcher and graph-API decoder fa - Network accessline 132
`SupportedArchitecture` and generated pipeline selection. - [registration_graph_api.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_graph_api.py): task/backend dispatcher and graph-API decoder fa - Network accessline 133
- [registration_graph_api.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_graph_api.py): - Network accessline 133
- [registration_graph_api.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_graph_api.py): - Network accessline 133
- [registration_graph_api.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_graph_api.py): - Network accessline 134
task/backend dispatcher and graph-API decoder factory. - [module_registration.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/module_registration.py): ModuleV3 role validation, vision, DP, LoRA, config finali - Network accessline 134
task/backend dispatcher and graph-API decoder factory. - [module_registration.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/module_registration.py): ModuleV3 role validation, vision, DP, LoRA, config finali - Network accessline 134
task/backend dispatcher and graph-API decoder factory. - [module_registration.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/module_registration.py): ModuleV3 role validation, vision, DP, LoRA, config finali - Network accessline 135
- [module_registration.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/module_registration.py): - Network accessline 135
- [module_registration.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/module_registration.py): - Network accessline 135
- [module_registration.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/module_registration.py): - Network accessline 136
ModuleV3 role validation, vision, DP, LoRA, config finalization and state binding. - [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): - Network accessline 136
ModuleV3 role validation, vision, DP, LoRA, config finalization and state binding. - [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): - Network accessline 136
ModuleV3 role validation, vision, DP, LoRA, config finalization and state binding. - [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): - Network accessline 137
binding. - [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): graph-API encoder contract and batching layouts. - Network accessline 137
binding. - [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): graph-API encoder contract and batching layouts. - Network accessline 137
binding. - [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): graph-API encoder contract and batching layouts. - Network accessline 138
- [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): - Network accessline 138
- [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): - Network accessline 138
- [registration_encoder.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_encoder.py): - Network accessline 139
graph-API encoder contract and batching layouts. - [role_contracts.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/role_contracts.py), [request_state.py](https://github.com/modular/modular/blob/main/max/pytho - Network accessline 139
graph-API encoder contract and batching layouts. - [role_contracts.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/role_contracts.py), [request_state.py](https://github.com/modular/modular/blob/main/max/pytho - Network accessline 139
graph-API encoder contract and batching layouts. - [role_contracts.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/role_contracts.py), [request_state.py](https://github.com/modular/modular/blob/main/max/pytho - Network accessline 140
- [role_contracts.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/role_contracts.py), - Network accessline 140
- [role_contracts.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/role_contracts.py), - Network accessline 140
- [role_contracts.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/role_contracts.py), - Network accessline 141
[request_state.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/request_state.py) - Network accessline 141
[request_state.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/request_state.py) - Network accessline 141
[request_state.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/request_state.py) - Network accessline 142
and [registration_runtime.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_runtime.py): decoder inputs/outputs, state lifecycle and deployment hooks. - Network accessline 142
and [registration_runtime.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_runtime.py): decoder inputs/outputs, state lifecycle and deployment hooks. - Network accessline 142
and [registration_runtime.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_runtime.py): decoder inputs/outputs, state lifecycle and deployment hooks. - Network accessline 143
[registration_runtime.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_runtime.py): - Network accessline 143
[registration_runtime.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_runtime.py): - Network accessline 143
[registration_runtime.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/registration_runtime.py): - Network accessline 144
decoder inputs/outputs, state lifecycle and deployment hooks. - [graph_specs.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/graph_specs.py): role names, including framework graph roles. - Network accessline 144
decoder inputs/outputs, state lifecycle and deployment hooks. - [graph_specs.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/graph_specs.py): role names, including framework graph roles. - Network accessline 144
decoder inputs/outputs, state lifecycle and deployment hooks. - [graph_specs.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/graph_specs.py): role names, including framework graph roles. - Network accessline 145
- [graph_specs.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/graph_specs.py): - Network accessline 145
- [graph_specs.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/graph_specs.py): - Network accessline 145
- [graph_specs.py](https://github.com/modular/modular/blob/main/max/python/max/pipelines/lib/graph_specs.py):
mojo-python-interop/SKILL.md
new-modular-project/SKILL.md
- Dangerous shell operationline 44
| Ubuntu/Debian | `sudo apt install gcc` | - Dangerous shell operationline 44
| Ubuntu/Debian | `sudo apt install gcc` | - Dangerous shell operationline 45
| Fedora/RHEL | `sudo dnf install gcc` | - Dangerous shell operationline 45
| Fedora/RHEL | `sudo dnf install gcc` | - Network accessline 49
**Windows users**: Mojo doesn't run natively on Windows. Install [WSL2](https://learn.microsoft.com/en-us/windows/wsl/install) - Network accessline 49
**Windows users**: Mojo doesn't run natively on Windows. Install [WSL2](https://learn.microsoft.com/en-us/windows/wsl/install) - Network accessline 50
Install [WSL2](https://learn.microsoft.com/en-us/windows/wsl/install) - Network accessline 50
Install [WSL2](https://learn.microsoft.com/en-us/windows/wsl/install) - Network accessline 61
First, determine whether `pixi` is installed. If it isn't available at the command line, install it using the latest instructions at <https://pixi.prefix.dev/latest/#installation>. - Network accessline 61
First, determine whether `pixi` is installed. If it isn't available at the command line, install it using the latest instructions at <https://pixi.prefix.dev/latest/#installation>. - Network accessline 63
<https://pixi.prefix.dev/latest/#installation>. - Network accessline 63
<https://pixi.prefix.dev/latest/#installation>. - Network accessline 70
# New project pixi init [PROJECT] \ -c https://conda.modular.com/max-nightly/ -c conda-forge \ - Network accessline 70
# New project pixi init [PROJECT] \ -c https://conda.modular.com/max-nightly/ -c conda-forge \ - Network accessline 71
pixi init [PROJECT] -c https://conda.modular.com/max-nightly/ -c conda-forge - Network accessline 71
pixi init [PROJECT] -c https://conda.modular.com/max-nightly/ -c conda-forge - Network accessline 72
-c https://conda.modular.com/max-nightly/ -c conda-forge \ - Network accessline 72
-c https://conda.modular.com/max-nightly/ -c conda-forge \ - Network accessline 77
# Existing project - add to pixi.toml channels first: # [workspace] # channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 77
# Existing project - add to pixi.toml channels first: # [workspace] # channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 78
# [workspace] # channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] pixi add [max / mojo] - Network accessline 78
# [workspace] # channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] pixi add [max / mojo] - Network accessline 79
# channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 79
# channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 86
# New project pixi init [PROJECT] \ -c https://conda.modular.com/max/ -c conda-forge \ - Network accessline 86
# New project pixi init [PROJECT] \ -c https://conda.modular.com/max/ -c conda-forge \ - Network accessline 87
pixi init [PROJECT] -c https://conda.modular.com/max/ -c conda-forge - Network accessline 87
pixi init [PROJECT] -c https://conda.modular.com/max/ -c conda-forge - Network accessline 88
-c https://conda.modular.com/max/ -c conda-forge \ - Network accessline 88
-c https://conda.modular.com/max/ -c conda-forge \ - Network accessline 117
uv init [PROJECT] && cd [PROJECT] uv add [max / mojo] \ --index https://whl.modular.com/nightly/simple/ \ - Network accessline 117
uv init [PROJECT] && cd [PROJECT] uv add [max / mojo] \ --index https://whl.modular.com/nightly/simple/ \ - Network accessline 118
uv add [max / mojo] --index https://whl.modular.com/nightly/simple/ - Network accessline 118
uv add [max / mojo] --index https://whl.modular.com/nightly/simple/ - Network accessline 119
--index https://whl.modular.com/nightly/simple/ \ - Network accessline 119
--index https://whl.modular.com/nightly/simple/ \ - Network accessline 137
uv venv uv pip install [max / mojo] \ --index https://whl.modular.com/nightly/simple/ \ - Network accessline 137
uv venv uv pip install [max / mojo] \ --index https://whl.modular.com/nightly/simple/ \ - Network accessline 138
uv pip install [max / mojo] --index https://whl.modular.com/nightly/simple/ - Network accessline 138
uv pip install [max / mojo] --index https://whl.modular.com/nightly/simple/ - Network accessline 139
--index https://whl.modular.com/nightly/simple/ \ - Network accessline 139
--index https://whl.modular.com/nightly/simple/ \ - Network accessline 167
python3 -m venv .venv && source .venv/bin/activate pip install --pre [max / mojo] \ --extra-index-url https://whl.modular.com/nightly/simple/ - Network accessline 167
python3 -m venv .venv && source .venv/bin/activate pip install --pre [max / mojo] \ --extra-index-url https://whl.modular.com/nightly/simple/ - Network accessline 168
pip install --pre [max / mojo] --extra-index-url https://whl.modular.com/nightly/simple/ - Network accessline 168
pip install --pre [max / mojo] --extra-index-url https://whl.modular.com/nightly/simple/ - Network accessline 169
--extra-index-url https://whl.modular.com/nightly/simple/ - Network accessline 169
--extra-index-url https://whl.modular.com/nightly/simple/ - Network accessline 195
conda install -c conda-forge \ -c https://conda.modular.com/max-nightly/ [max / mojo] - Network accessline 195
conda install -c conda-forge \ -c https://conda.modular.com/max-nightly/ [max / mojo] - Network accessline 196
-c https://conda.modular.com/max-nightly/ [max / mojo] - Network accessline 196
-c https://conda.modular.com/max-nightly/ [max / mojo] - Network accessline 202
conda install -c conda-forge \ -c https://conda.modular.com/max/ [max / mojo] - Network accessline 202
conda install -c conda-forge \ -c https://conda.modular.com/max/ [max / mojo] - Network accessline 203
-c https://conda.modular.com/max/ [max / mojo] - Network accessline 203
-c https://conda.modular.com/max/ [max / mojo] - Network accessline 238
- [Mojo Installation Guide](https://mojolang.org/install) - Network accessline 238
- [Mojo Installation Guide](https://mojolang.org/install) - Network accessline 239
- [Mojo Stable Docs](https://mojolang.org/docs/) - Network accessline 239
- [Mojo Stable Docs](https://mojolang.org/docs/) - Network accessline 240
- [Mojo Nightly Docs](https://mojolang.org/nightly/docs/) - Network accessline 240
- [Mojo Nightly Docs](https://mojolang.org/nightly/docs/)
profile-model/SKILL.md
- Network accessline 150
pixi init quickstart -c https://conda.modular.com/max-nightly/ -c conda-forge - Network accessline 150
pixi init quickstart -c https://conda.modular.com/max-nightly/ -c conda-forge - Dangerous shell operationline 171
`sudo apt-get install nsight-systems`. - Dangerous shell operationline 171
`sudo apt-get install nsight-systems`.
profile-model/references/nsys.md
- Dangerous shell operationline 10
`sudo apt-get install nsight-systems` or download from NVIDIA. Use version - Dangerous shell operationline 10
`sudo apt-get install nsight-systems` or download from NVIDIA. Use version - Network accessline 103
2. Wait for `🚀 Server ready on http://0.0.0.0:8000`. (Fail fast: if the - Network accessline 103
2. Wait for `🚀 Server ready on http://0.0.0.0:8000`. (Fail fast: if the
profile-model/references/rocprofv3.md
- Network accessline 4
MI355, and similar). `rocprofv3` captures a kernel trace plus marker (NVTX-equivalent) trace and writes a Perfetto `.pftrace` you open at <https://ui.perfetto.dev>. - Network accessline 4
MI355, and similar). `rocprofv3` captures a kernel trace plus marker (NVTX-equivalent) trace and writes a Perfetto `.pftrace` you open at <https://ui.perfetto.dev>. - Network accessline 6
<https://ui.perfetto.dev>. - Network accessline 6
<https://ui.perfetto.dev>. - Network accessline 72
A valid Perfetto file is non-trivial in size and begins with the Perfetto magic header. Report every `.pftrace` path and size, and tell the user to open it at <https://ui.perfetto.dev> (drag-and-drop) — the marker track shows the - Network accessline 72
A valid Perfetto file is non-trivial in size and begins with the Perfetto magic header. Report every `.pftrace` path and size, and tell the user to open it at <https://ui.perfetto.dev> (drag-and-drop) — the marker track shows the - Network accessline 73
magic header. Report every `.pftrace` path and size, and tell the user to open it at <https://ui.perfetto.dev> (drag-and-drop) — the marker track shows the phase labels over the kernel timeline. - Network accessline 73
magic header. Report every `.pftrace` path and size, and tell the user to open it at <https://ui.perfetto.dev> (drag-and-drop) — the marker track shows the phase labels over the kernel timeline. - Network accessline 74
it at <https://ui.perfetto.dev> (drag-and-drop) — the marker track shows the - Network accessline 74
it at <https://ui.perfetto.dev> (drag-and-drop) — the marker track shows the
profile-model/scripts/gpu_snapshot.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
serve-model/SKILL.md
- Network accessline 71
# 4. Confirm with one request (model field must equal --served-model-name): curl -s http://localhost:8000/v1/chat/completions -H "Content-Type: application/json" \ - Network accessline 71
# 4. Confirm with one request (model field must equal --served-model-name): curl -s http://localhost:8000/v1/chat/completions -H "Content-Type: application/json" \ - Network accessline 72
curl -s http://localhost:8000/v1/chat/completions -H "Content-Type: application/json" \ - Network accessline 72
curl -s http://localhost:8000/v1/chat/completions -H "Content-Type: application/json" \ - Network accessline 94
If the `max serve` command isn't available, set up an environment. **pixi** is the default; the key detail is the conda channel `https://conda.modular.com/max-nightly/` plus `conda-forge`. If `pixi` isn't - Network accessline 94
If the `max serve` command isn't available, set up an environment. **pixi** is the default; the key detail is the conda channel `https://conda.modular.com/max-nightly/` plus `conda-forge`. If `pixi` isn't - Network accessline 95
the default; the key detail is the conda channel `https://conda.modular.com/max-nightly/` plus `conda-forge`. If `pixi` isn't already installed, install it using the latest instructions at - Network accessline 95
the default; the key detail is the conda channel `https://conda.modular.com/max-nightly/` plus `conda-forge`. If `pixi` isn't already installed, install it using the latest instructions at - Network accessline 96
`https://conda.modular.com/max-nightly/` plus `conda-forge`. If `pixi` isn't - Network accessline 96
`https://conda.modular.com/max-nightly/` plus `conda-forge`. If `pixi` isn't - Network accessline 98
<https://pixi.prefix.dev/latest/#installation>, then: - Network accessline 98
<https://pixi.prefix.dev/latest/#installation>, then: - Network accessline 101
# pixi (conda channels) pixi init my-max-project \ -c https://conda.modular.com/max-nightly/ -c conda-forge && cd my-max-project - Network accessline 101
# pixi (conda channels) pixi init my-max-project \ -c https://conda.modular.com/max-nightly/ -c conda-forge && cd my-max-project - Network accessline 102
pixi init my-max-project -c https://conda.modular.com/max-nightly/ -c conda-forge && cd my-max-project - Network accessline 102
pixi init my-max-project -c https://conda.modular.com/max-nightly/ -c conda-forge && cd my-max-project - Network accessline 103
-c https://conda.modular.com/max-nightly/ -c conda-forge && cd my-max-project - Network accessline 103
-c https://conda.modular.com/max-nightly/ -c conda-forge && cd my-max-project - Network accessline 110
[workspace] # or [project] on older pixi channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 110
[workspace] # or [project] on older pixi channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 111
channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 111
channels = ["https://conda.modular.com/max-nightly/", "conda-forge"] - Network accessline 114
Some users prefer **uv**, which pulls MAX from Modular's wheel index instead of conda. If `uv` isn't already installed, install it using the latest instructions at <https://docs.astral.sh/uv/getting-started/installation/>, - Network accessline 114
Some users prefer **uv**, which pulls MAX from Modular's wheel index instead of conda. If `uv` isn't already installed, install it using the latest instructions at <https://docs.astral.sh/uv/getting-started/installation/>, - Network accessline 115
conda. If `uv` isn't already installed, install it using the latest instructions at <https://docs.astral.sh/uv/getting-started/installation/>, then: - Network accessline 115
conda. If `uv` isn't already installed, install it using the latest instructions at <https://docs.astral.sh/uv/getting-started/installation/>, then: - Network accessline 116
instructions at <https://docs.astral.sh/uv/getting-started/installation/>, - Network accessline 116
instructions at <https://docs.astral.sh/uv/getting-started/installation/>, - Network accessline 121
uv init my-max-project && cd my-max-project uv venv && source .venv/bin/activate uv add "max[serve]" --index https://whl.modular.com/nightly/simple/ --prerelease allow - Network accessline 121
uv init my-max-project && cd my-max-project uv venv && source .venv/bin/activate uv add "max[serve]" --index https://whl.modular.com/nightly/simple/ --prerelease allow - Network accessline 123
uv add "max[serve]" --index https://whl.modular.com/nightly/simple/ --prerelease allow - Network accessline 123
uv add "max[serve]" --index https://whl.modular.com/nightly/simple/ --prerelease allow - Network accessline 271
Server ready on http://0.0.0.0:8000 (Press CTRL+C to quit) - Network accessline 271
Server ready on http://0.0.0.0:8000 (Press CTRL+C to quit) - Network accessline 283
curl -s http://localhost:8000/v1/health # 200 when ready - Network accessline 283
curl -s http://localhost:8000/v1/health # 200 when ready - Network accessline 284
curl -s http://localhost:8000/v1/models # served model name - Network accessline 284
curl -s http://localhost:8000/v1/models # served model name - Network accessline 286
curl -X POST http://localhost:8000/v1/chat/completions \ - Network accessline 286
curl -X POST http://localhost:8000/v1/chat/completions \
serve-model/references/troubleshooting.md
- Network accessline 34
curl -s http://localhost:8000/v1/health # 200 when ready - Network accessline 34
curl -s http://localhost:8000/v1/health # 200 when ready - Network accessline 35
curl -s http://localhost:8000/v1/models # lists the served model name - Network accessline 35
curl -s http://localhost:8000/v1/models # lists the served model name
serve-model/scripts/suggest_serve_command.py
- Executable scriptline 1
Executable file - Executable scriptline 1
Executable file - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 3
# # Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 4
# Licensed under the Apache License v2.0 with LLVM Exceptions: # https://llvm.org/LICENSE.txt # - Network accessline 5
# https://llvm.org/LICENSE.txt - Network accessline 5
# https://llvm.org/LICENSE.txt
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Some files could not be read; the facts below still hold.
Signal confidence: 500 facts
- Network access in code of the SKILL.mdbenchmark-model/SKILL.md:58High confidence
- Network access in code of the SKILL.mdbenchmark-model/SKILL.md:59High confidence
- Network access in code of the SKILL.mdbenchmark-model/SKILL.md:120High confidence
- Network access in code of the SKILL.mdbenchmark-model/SKILL.md:131High confidence
- Network access in prose of the SKILL.mdbenchmark-model/SKILL.md:168Low confidence
- Network access in prose of the SKILL.mdbenchmark-model/SKILL.md:169Low confidence
- Network access in prose of the SKILL.mdbenchmark-model/SKILL.md:170Low confidence
- Network access in prose of documentationbenchmark-model/references/flags.md:22Low confidence
- Network access in prose of documentationbenchmark-model/references/flags.md:23Low confidence
- Network access in prose of documentationbenchmark-model/references/flags.md:24Low confidence
- Network access in prose of documentationbenchmark-model/references/troubleshooting.md:4Low confidence
- Network access in prose of documentationbenchmark-model/references/troubleshooting.md:5Low confidence
- Network access in prose of documentationbenchmark-model/references/troubleshooting.md:6Low confidence
- Network access in prose of documentationbenchmark-model/references/troubleshooting.md:12Low confidence
- Network access in prose of documentationbenchmark-model/references/troubleshooting.md:13Low confidence
- Network access in prose of documentationbenchmark-model/references/troubleshooting.md:14Low confidence
- Network access in prose of documentationbenchmark-model/references/troubleshooting.md:15Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:46Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:46Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:46Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:47Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:47Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:47Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:48Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:48Low confidence
- Network access in prose of the SKILL.mddebug-model/SKILL.md:48Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:27Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:27Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:93Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:93Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:176Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:176Low confidence
- Network access in prose of documentationdebug-model/references/comparator-build.md:218Low confidence
- Network access in prose of documentationdebug-model/references/comparator-build.md:218Low confidence
- Network access in prose of documentationdebug-model/references/comparator-build.md:220Low confidence
- Network access in prose of documentationdebug-model/references/comparator-build.md:220Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:227Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:227Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:228Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valuedebug-model/references/comparator-build.md:228Low confidence
- Network access in prose of the SKILL.mdeval-model/SKILL.md:313Low confidence
- Network access in prose of the SKILL.mdeval-model/SKILL.md:313Low confidence
- Network access in code of documentationeval-model/references/examples.md:135Low confidence
- Network access in code of documentationeval-model/references/examples.md:135Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valueeval-model/references/examples.md:135Low confidence
- Access to secrets in code of documentation: a variable, placeholder or call, not a stored valueeval-model/references/examples.md:135Low confidence
- Network access in code of documentationeval-model/references/examples.md:136Low confidence
- Network access in code of documentationeval-model/references/examples.md:136Low confidence
- Network access in code of documentationeval-model/references/examples.md:137Low confidence
- Network access in code of documentationeval-model/references/examples.md:137Low confidence
and 450 more
Only the first 500 facts are listed.
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Manifest .claude-plugin/marketplace.json | .claude/skills/<skill-name>/ |
| Codex | Compatible | Standard SKILL.md benchmark-model/SKILL.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md benchmark-model/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md benchmark-model/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md benchmark-model/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md benchmark-model/SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.
