Menu

Not in the top 1,000 by GitHub stars

danielmeppiel/genesis

Markdown that steers an LLM is code. Genesis is the architectural layer for designing multi-agent, multi-skill systems -- with named patterns, contracts, and substrate portability, before you write them.

Engineering

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

dev/skills/genesis-evals/SKILL.md

Week of Nov 2, 2026 · +0 / −0 lines

Before Local scripts → After Local scripts

18e73c35e5ce → 18e73c35e5ce

Compare on GitHub

Week of Oct 26, 2026 · Line counts unavailable

Before Not checked yet → After Local scripts

First observation → 18e73c35e5ce

No earlier version was observed for comparison.

Compare on GitHub

skills/genesis/SKILL.md

Week of Nov 2, 2026 · +0 / −0 lines

Before Network or broad shell → After Network or broad shell

18e73c35e5ce → 18e73c35e5ce

Compare on GitHub

Week of Oct 26, 2026 · Line counts unavailable

Before Not checked yet → After Network or broad shell

First observation → 18e73c35e5ce

No earlier version was observed for comparison.

Compare on GitHub

View on GitHubAdd to my skills

Verified commit 18e73c35e5cefaf431fbe4e080e8f125ebfece33

78GitHub stars
1star in 7 days
2skills in this repository
0commits in 30 days

Heat

1.8
  • 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

Oct 25, 2026Nov 5, 202678

30-day change not yet available: not enough history.

Individual skills in this repository

How it works

genesis-evals

When your agent uses it

Use this skill to run the genesis maintainer-side eval suite against a target model (default: claude-opus-4.7). Activate when validating a genesis PR, when changing the genesis catalogue (architectural-patterns, primitives, design-patterns, refactor-patterns, composition-substrate, pattern-tradeoffs, SKILL.md), or when the operator asks to "run evals" or "regenerate the eval matrix". This skill orchestrates parallel cold sub-agent spawns via the harness's task tool, scores deterministically, and converges P>=0.8 / N>=0.8 / R==1.0 within max 3 iteration loops. This skill is contributor-only -- it lives under dev/skills/ (OUTSIDE .apm/) and is NOT shipped inside the user-facing skills/genesis/ bundle (BUNDLE LEAKAGE discipline). See "Why this lives outside .apm/" below.

What it loads into context

Frontmatter

217 tokens

Always available to the agent

SKILL.md

2,561 tokens

When the skill is activated

Reference documents

0 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

  1. Why this lives outside `.apm/`
  2. When to activate
  3. Hard rules
  4. Process
  5. Step 1 -- validate
  6. Step 2 -- mint run-id
  7. Step 3 -- fan-out spawns
  8. Step 4 -- score
  9. Step 5 -- read summary
  10. Step 6 -- decide convergence
  11. Step 7 -- iterate (cap = 3)
  12. Updating a PR description with the matrix
  13. Adding a new scenario
  14. Gate-authoring tips (lessons from RUN_ID=20260426T104629Z)
  15. Retiring a scenario
  16. Setup (one-time)

Scripts executed separately

  • scripts/score_run.py
  • scripts/spawn_record.py
  • scripts/validate_scenarios.py

Files

  • .gitignoreasset7 bytes
  • SKILL.mdcore9,935 bytes
  • SKILL.mdcore9,935 bytes
  • SKILL.mdcore9,935 bytes
  • SKILL.mdcore9,935 bytes
  • SKILL.mdcore9,935 bytes
  • SKILL.mdcore9,935 bytes
  • requirements.txtdoc28 bytes
  • runs/.gitignoreasset29 bytes
  • scenarios/n-001-debug-bug.ymlasset1,012 bytes
  • scenarios/n-002-write-doc.ymlasset895 bytes
  • scenarios/n-003-explain-concept.ymlasset619 bytes
  • scenarios/n-004-cost-near-miss.ymlasset1,008 bytes
  • scenarios/p-001-a10-recognition.ymlasset874 bytes
  • scenarios/p-002-soc-extraction.ymlasset1,086 bytes
  • scenarios/p-003-panel-fanout.ymlasset908 bytes
  • scenarios/p-004-a12-gradient-workflow.ymlasset1,131 bytes
  • scenarios/p-005-b13-cache-invalidator.ymlasset1,172 bytes
  • scenarios/p-006-r5-cost-prune.ymlasset991 bytes
  • scenarios/p-007-cost-stance-projection.ymlasset1,019 bytes
  • scenarios/p-008-b12-model-router.ymlasset1,029 bytes
  • scenarios/r-001-substrate-validation.ymlasset1,022 bytes
  • scenarios/r-002-substrate-implied-not-github.ymlasset936 bytes
  • schema/scenario.schema.jsonasset2,480 bytes
  • scripts/score_run.pyruns5,769 bytes
  • scripts/spawn_record.pyruns2,748 bytes
  • scripts/validate_scenarios.pyruns2,211 bytes
Read SKILL.md

# genesis-evals: maintainer-side eval runner

Run the genesis self-eval suite. Steers the parent LLM session to
orchestrate cold sub-agent spawns, capture responses, score
deterministically, and report convergence.

## Why this lives outside `.apm/`

Genesis ships to USERS via npx / `apm install`. Eval scenarios LOOK
LIKE real user requests (that is the point). Colocating them under
`skills/genesis/evals/` would risk DISPATCH CONTAMINATION (an
over-eager harness loader pulling scenario prompts into the active
context) and PAYLOAD BLOAT for users who never run evals.
View the complete source on GitHub

Skills included

  • genesis-evals
  • genesis

Repository facts

License
Apache-2.0
Created
Apr 25, 2026
Last pushed
May 30, 2026
Forks
11
Open issues
3
Project activity
Slowing

Before you install

Network or broad shell

Automated signals, not a guarantee

Checked Nov 6, 2026

Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.

Show 8 notes
  • dev/skills/genesis-evals/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • dev/skills/genesis-evals/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • dev/skills/genesis-evals/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • dev/skills/genesis-evals/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • dev/skills/genesis-evals/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/genesis/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/genesis/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/genesis/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
Show all 102 signals by file

dev/skills/genesis-evals/schema/scenario.schema.json

  • Network accessline 1
    {   "$schema": "http://json-schema.org/draft-07/schema#",
  • Network accessline 2
    "$schema": "http://json-schema.org/draft-07/schema#",

dev/skills/genesis-evals/scripts/score_run.py

  • Executable scriptline 1
    #!/usr/bin/env python3

dev/skills/genesis-evals/scripts/spawn_record.py

  • Executable scriptline 1
    #!/usr/bin/env python3

dev/skills/genesis-evals/scripts/validate_scenarios.py

  • Executable scriptline 1
    #!/usr/bin/env python3

skills/genesis/assets/design-patterns.md

  • Dangerous shell operationline 1119
    - security warnings (do not telegraph "rm -rf safe")

skills/genesis/assets/module-system-adapters/apm.md

  • Network accessline 32
    - package layout (`.apm/` source root, `dev/` for maintainer   assets) -- canonical reference:   https://danielmeppiel.github.io/awd-cli/introduction/anatomy-of-an-apm-package/
  • Network accessline 34
    https://danielmeppiel.github.io/awd-cli/introduction/anatomy-of-an-apm-package/

skills/genesis/assets/primitives.md

  • Network accessline 38
    INDUSTRY TERMS: "skill" ([agentskills.io](https://agentskills.io)),
  • Network accessline 69
    - https://agentskills.io/skill-creation/best-practices
  • Network accessline 71
    prescriptiveness, procedures over declarations, refine-with-      execution, real-expertise sourcing. - https://agentskills.io/skill-creation/optimizing-descriptions
  • Network accessline 72
    execution, real-expertise sourcing. - https://agentskills.io/skill-creation/optimizing-descriptions   -> imperative phrasing, user-intent framing, indirect-trigger
  • Network accessline 73
    - https://agentskills.io/skill-creation/optimizing-descriptions
  • Network accessline 75
    ("be pushy") clauses, 1024-character hard cap on `description`,      trigger-eval split (~20 queries 60/40 train/val). - https://agentskills.io/skill-creation/evaluating-skills
  • Network accessline 76
    trigger-eval split (~20 queries 60/40 train/val). - https://agentskills.io/skill-creation/evaluating-skills   -> `evals/evals.json` schema, with-skill vs without-skill baseline,
  • Network accessline 77
    - https://agentskills.io/skill-creation/evaluating-skills
  • Network accessline 78
    -> `evals/evals.json` schema, with-skill vs without-skill baseline,      iteration workspace, when assertions land. - https://agentskills.io/skill-creation/using-scripts
  • Network accessline 79
    iteration workspace, when assertions land. - https://agentskills.io/skill-creation/using-scripts   -> `scripts/` directory conventions, version pinning,
  • Network accessline 80
    - https://agentskills.io/skill-creation/using-scripts
  • Network accessline 82
    non-interactive shell requirement, `--help` doc, structured      stdout vs diagnostic stderr. - https://agentskills.io/specification
  • Network accessline 83
    stdout vs diagnostic stderr. - https://agentskills.io/specification   -> `name` regex (1-64 chars, `[a-z0-9-]`, must equal parent dir),
  • Network accessline 84
    - https://agentskills.io/specification
  • Network accessline 230
    Each primitive earns its keep against PROSE ([danielmeppiel.github.io/awesome-ai-native](https://danielmeppiel.github.io/awesome-ai-native/)):
  • Network accessline 231
    ([danielmeppiel.github.io/awesome-ai-native](https://danielmeppiel.github.io/awesome-ai-native/)):
  • Network accessline 293
    CANONICAL EXTERNAL CORPUS, BOUNDED. The Model Context Protocol ([modelcontextprotocol.io](https://modelcontextprotocol.io)) is the
  • Network accessline 294
    ([modelcontextprotocol.io](https://modelcontextprotocol.io)) is the

skills/genesis/assets/runtime-affordances/model-catalog.md

  • Network accessline 285
    - Azure Foundry reasoning docs (effort levels, defaults, SKU   matrix): https://learn.microsoft.com/azure/foundry/openai/how-to/reasoning
  • Network accessline 286
    matrix): https://learn.microsoft.com/azure/foundry/openai/how-to/reasoning
  • Network accessline 287
    - Azure Foundry responses API (ReasoningEffort component): https://learn.microsoft.com/rest/api/aifoundry/azureopenai/responses
  • Network accessline 288
    - OpenAI GPT-5 prompting guide (referenced by Azure docs): https://cookbook.openai.com/examples/gpt-5/gpt-5_prompting_guide

skills/genesis/assets/runtime-affordances/per-harness/claude-code.md

  • Network accessline 7
    Official docs cited: - https://docs.claude.com/en/docs/claude-code/overview
  • Network accessline 8
    - https://docs.claude.com/en/docs/claude-code/overview
  • Network accessline 9
    - https://docs.claude.com/en/docs/claude-code/sub-agents
  • Network accessline 10
    - https://docs.claude.com/en/docs/claude-code/skills
  • Network accessline 11
    - https://docs.claude.com/en/docs/claude-code/memory
  • Network accessline 12
    - https://docs.claude.com/en/docs/claude-code/settings
  • Network accessline 13
    - https://docs.claude.com/en/docs/claude-code/hooks
  • Network accessline 23
    Claude Code model default is claude-opus-4-1. Instructions field mirrors persona   body concept. - Source: https://docs.claude.com/en/docs/claude-code/sub-agents
  • Network accessline 25
    - Source: https://docs.claude.com/en/docs/claude-code/sub-agents
  • Network accessline 36
    - Notes: aligns with agentskills.io registry standard; description-driven matching is the   primary trigger; skills are loaded into context when recommended - Source: https://docs.claude.com/en/docs/claude-code/skills
  • Network accessline 38
    - Source: https://docs.claude.com/en/docs/claude-code/skills
  • Network accessline 48
    can import external markdown via @path syntax within CLAUDE.md body;   child CLAUDE.md files override parent rules in same directory subtree - Source: https://docs.claude.com/en/docs/claude-code/memory
  • Network accessline 50
    - Source: https://docs.claude.com/en/docs/claude-code/memory
  • Network accessline 61
    final response as text returned from Task call; subagent_type parameter is   the subagent name (without file extension) - Source: https://docs.claude.com/en/docs/claude-code/sub-agents
  • Network accessline 63
    - Source: https://docs.claude.com/en/docs/claude-code/sub-agents
  • Network accessline 75
    - Notes: hooks run within current session context; not suitable for cross-session   persistence; each hook invocation receives context and can return modifications - Source: https://docs.claude.com/en/docs/claude-code/hooks
  • Network accessline 77
    - Source: https://docs.claude.com/en/docs/claude-code/hooks
  • Network accessline 149
    Source: https://www.anthropic.com/pricing#api

skills/genesis/assets/runtime-affordances/per-harness/codex.md

  • Network accessline 7
    Official docs cited: - https://github.com/openai/codex
  • Network accessline 8
    - https://github.com/openai/codex
  • Network accessline 9
    - https://developers.openai.com/docs
  • Network accessline 16
    - Notes: Codex reads AGENTS.md wholesale; personas are absorbed into   its narrative structure, not isolated. No persona-specific fallback. - Source: https://github.com/openai/codex (no discrete persona format)
  • Network accessline 18
    - Source: https://github.com/openai/codex (no discrete persona format)
  • Network accessline 122
    Token pass-through. Per-Mtok rates (OpenAI public API, verified 2025-11-14): consult https://openai.com/api/pricing for the live
  • Network accessline 123
    2025-11-14): consult https://openai.com/api/pricing for the live

skills/genesis/assets/runtime-affordances/per-harness/copilot.md

  • Network accessline 7
    Official docs cited: - https://docs.github.com/en/copilot/reference/custom-agents-configuration
  • Network accessline 8
    - https://docs.github.com/en/copilot/reference/custom-agents-configuration
  • Network accessline 9
    (CANONICAL custom agent frontmatter spec -- `model`, `tools`, `target`,   `disable-model-invocation`, `user-invocable`, `mcp-servers`) - https://docs.github.com/en/copilot/how-tos/copilot-cli/customize-copilot/create-custom-agents-for-cli
  • Network accessline 10
    `disable-model-invocation`, `user-invocable`, `mcp-servers`) - https://docs.github.com/en/copilot/how-tos/copilot-cli/customize-copilot/create-custom-agents-for-cli - https://docs.github.com/en/copilot/concepts/agents/coding-agent/about-hoo
  • Network accessline 11
    - https://docs.github.com/en/copilot/how-tos/copilot-cli/customize-copilot/create-custom-agents-for-cli
  • Network accessline 12
    - https://docs.github.com/en/copilot/concepts/agents/coding-agent/about-hooks
  • Network accessline 52
    `.md` or `.agent.md`); lowest-level config wins (repo > org > enterprise).   Prompt body capped at 30,000 characters. - Source: https://docs.github.com/en/copilot/reference/custom-agents-configuration
  • Network accessline 54
    - Source: https://docs.github.com/en/copilot/reference/custom-agents-configuration
  • Network accessline 77
    deployed as SKILL.md at skill root (not renamed). Aligns with   agentskills.io registry contract; description is the primary search key. - Source: https://docs.github.com/en/copilot/customizing-copilot/adding-custom-instructions-for-github-
  • Network accessline 79
    - Source: https://docs.github.com/en/copilot/customizing-copilot/adding-custom-instructions-for-github-copilot
  • Network accessline 89
    work path matches the glob. At user scope, instructions are visible to   all projects. Pattern matching happens per-thread at runtime. - Source: https://docs.github.com/en/copilot/customizing-copilot/adding-custom-instructions-for-github-co
  • Network accessline 91
    - Source: https://docs.github.com/en/copilot/customizing-copilot/adding-custom-instructions-for-github-copilot
  • Network accessline 185
    Source: https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing

skills/genesis/assets/runtime-affordances/per-harness/cursor.md

  • Network accessline 6
    Official docs cited: - https://docs.cursor.com/context/rules
  • Network accessline 7
    - https://docs.cursor.com/context/rules
  • Network accessline 8
    - https://docs.cursor.com/skills
  • Network accessline 9
    - https://docs.cursor.com/subagents
  • Network accessline 10
    - https://docs.cursor.com/hooks
  • Network accessline 21
    - Notes: alwaysApply=false means Cursor Agent decides relevance.   Nested AGENTS.md files are supported in subdirectories. - Source: https://docs.cursor.com/context/rules
  • Network accessline 23
    - Source: https://docs.cursor.com/context/rules
  • Network accessline 33
    - Notes: Cursor implements agentskills.io SKILL.md container and discovery;   assets load on demand as agents execute steps. - Source: https://docs.cursor.com/skills
  • Network accessline 35
    - Source: https://docs.cursor.com/skills
  • Network accessline 47
    - Legacy: AGENTS.md (plain markdown, no frontmatter, auto-loaded) - Notes: Cursor Settings UI also supports user-level rules (not file-based). - Source: https://docs.cursor.com/context/rules
  • Network accessline 49
    - Source: https://docs.cursor.com/context/rules
  • Network accessline 60
    - Notes: Subagents are agent-to-agent dispatch, not a task-level primitive.   No built-in programmatic spawn from within SKILL.md step execution. - Source: https://docs.cursor.com/subagents
  • Network accessline 62
    - Source: https://docs.cursor.com/subagents
  • Network accessline 71
    - Scope: Repo-local; not yet integrated with global ~/.cursor/ scope. - Notes: Hook semantics not fully documented in official API. - Source: https://docs.cursor.com/hooks
  • Network accessline 73
    - Source: https://docs.cursor.com/hooks
  • Network accessline 140
    Source: https://cursor.com/pricing (live).

skills/genesis/assets/runtime-affordances/per-harness/opencode.md

  • Network accessline 7
    Official docs cited: - https://github.com/anomalyco/opencode
  • Network accessline 8
    - https://github.com/anomalyco/opencode

skills/genesis/assets/runtime-affordances/per-trigger-surface/gh-aw.md

  • Network accessline 14
    Official docs cited (load-bearing for this adapter): - https://github.github.com/gh-aw/
  • Network accessline 15
    - https://github.github.com/gh-aw/
  • Network accessline 16
    -> entry point; project overview - https://github.github.com/gh-aw/introduction/architecture/   -> three-layer trust model (substrate / configuration / plan);
  • Network accessline 17
    - https://github.github.com/gh-aw/introduction/architecture/
  • Network accessline 19
    defines the SANDBOXING + CAPABILITY_GATING + AUDIT_SURFACE      guarantees this adapter maps to substrate primitive #5 - https://github.github.com/gh-aw/reference/frontmatter-full/
  • Network accessline 20
    guarantees this adapter maps to substrate primitive #5 - https://github.github.com/gh-aw/reference/frontmatter-full/   -> CANONICAL spec for every field this adapter mentions; do NOT
  • Network accessline 21
    - https://github.github.com/gh-aw/reference/frontmatter-full/
  • Network accessline 22
    -> CANONICAL spec for every field this adapter mentions; do NOT      redocument here, cite this page when emitting workflow markdown - https://github.github.com/gh-aw/reference/dependencies/
  • Network accessline 23
    redocument here, cite this page when emitting workflow markdown - https://github.github.com/gh-aw/reference/dependencies/   -> APM-as-dependencies story; the bridge between this trigger-
  • Network accessline 24
    - https://github.github.com/gh-aw/reference/dependencies/
  • Network accessline 25
    -> APM-as-dependencies story; the bridge between this trigger-      surface adapter and the existing `module-system-adapters/apm.md` - https://microsoft.github.io/apm/integrations/gh-aw/
  • Network accessline 26
    surface adapter and the existing `module-system-adapters/apm.md` - https://microsoft.github.io/apm/integrations/gh-aw/   -> APM-side view of the same integration; cite when designing
  • Network accessline 27
    - https://microsoft.github.io/apm/integrations/gh-aw/

skills/genesis/examples/05-pr-review-verdict.md

  • Network accessline 238
    DECLARATION MECHANISMS for runtime tools (NOT primitive deps, but loader-visible declarations to avoid PHANTOM DEPENDENCY):  - `gh`, `git`: preloaded by GitHub Actions runner image; no install needed; trigger orchestrator sets `permissions:
  • Dangerous shell operationline 240
    - `gitleaks`: trigger orchestrator setup step pins and installs (`curl -sSfL ... | sh`); lens-security PROBES at use-site (`command -v gitleaks` -> if missing, lens returns `secret_scan: unavailable` finding rather than fabricating).
  • Network accessline 240
    - `gitleaks`: trigger orchestrator setup step pins and installs (`curl -sSfL ... | sh`); lens-security PROBES at use-site (`command -v gitleaks` -> if missing, lens returns `secret_scan: unavailable` finding rather than fabricating).

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: 100 facts

and 50 more

Works with

Agent compatibility, from the files in this repository
AgentStatusEvidenceInstall to
Claude CodeCompatibleStandard SKILL.md dev/skills/genesis-evals/SKILL.md.claude/skills/<skill-name>/
CodexCompatibleStandard SKILL.md dev/skills/genesis-evals/SKILL.md.agents/skills/<skill-name>/
CursorCompatibleStandard SKILL.md dev/skills/genesis-evals/SKILL.md.cursor/skills/<skill-name>/
Gemini CLICompatibleStandard SKILL.md dev/skills/genesis-evals/SKILL.md.gemini/skills/<skill-name>/
GitHub CopilotCompatibleStandard SKILL.md dev/skills/genesis-evals/SKILL.md.github/skills/<skill-name>/
OpenCodeCompatibleStandard SKILL.md dev/skills/genesis-evals/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.

You might also use

NVIDIA/skills

Agent Skills for NVIDIA products — install into Claude Code, Codex, and other coding agents to run Physical AI, robotics, simulation, CUDA, and RAG workflows end to end.

3.9kGitHub stars

FrancyJGLisboa/agent-skills-platform

Build tested agent skills and govern their lifecycle through a user-defined marketplace: evidence, discovery, updates, rollback, quarantine, and 17-platform distribution.

2.4kGitHub stars

doccker/cc-use-exp

⛔ No longer maintained → dev-agent-kit/recipes 让 Claude Code、Antigravity、Gemini CLI、Codex、Cursor 开箱即用的分层配置模板，总结十多年的日常开发经验

1.1kGitHub stars

ericrisco/rsc-harness

Your agent invents things because it has no memory, and can't touch your database because it has no arms. rsc is the meta-harness that gives it both, plus the trade to know the jo…

525GitHub stars

nanaism/yomiyasu

AIが生成した日本語を推敲するスキル。利用者が yomiyasu の名前を出して頼んだとき（「/yomiyasu」、「$yomiyasu:yomiyasu」、「yomiyasu で推敲して」など）と、スキルを使った日本語の推敲…

2.9kGitHub stars

917Dhj/DeepPaperNote

DeepPaperNote is an agent skill for deep-reading a single paper and generating high-quality Obsidian-style research notes. Works with Claude Code, Codex, Cursor, Copilot, Gemini C…

1.3kGitHub stars