Not in the top 1,000 by GitHub stars
xingwudao/open-xquant
AI 量化研究框架:AI Agent 驱动策略回测、因子研究、稳健性检验、审计报告与实盘交易 | Agentic Quant Research Kernel
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
agent/skills/audit-artifact-lineage/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/audit-runtime-semantics/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/audit-strategy-idea/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/audit-strategy-spec/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/author-component/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/brainstorm-strategy-idea/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/build-report-charts/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/build-rule/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/build-strategy-spec/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/build-universe/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/compare-experiments/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/compare-strategy-versions/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/configure-trade-execution/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/create-component/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/create-indicator/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/create-portfolio-optimizer/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/create-rule/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/create-signal/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/evaluate-cross-sectional/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/evaluate-factor/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/evaluate-time-series/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/explore-data/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/govern-research-workspace/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/manage-live-trading/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/manage-strategy-version/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/monitor-strategy-run/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/open-xquant/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/plot-indicators/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/review-performance/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/review-research-report/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/run-authorized-backtest/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/screen-factors/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/select-final-version/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/tune-parameters/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
agent/skills/write-research-report/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → eba4ac6af740
No earlier version was observed for comparison.
Verified commit eba4ac6af7401e4b8bd4858873ec2892e03120d9
git clone --filter=blob:none --no-checkout https://github.com/xingwudao/open-xquant.git .hotskills/xingwudao-open-xquant && git -C .hotskills/xingwudao-open-xquant checkout --detach eba4ac6af7401e4b8bd4858873ec2892e03120d9
# Copy each selected skill directory from .hotskills/xingwudao-open-xquant into .claude/skills/<skill-name>/git clone --filter=blob:none --no-checkout https://github.com/xingwudao/open-xquant.git .hotskills/xingwudao-open-xquant && git -C .hotskills/xingwudao-open-xquant checkout --detach eba4ac6af7401e4b8bd4858873ec2892e03120d9
# Copy each selected skill directory from .hotskills/xingwudao-open-xquant into .agents/skills/<skill-name>/Heat
2.5- 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
- audit-artifact-lineage
agent/skills/audit-artifact-lineage - audit-runtime-semantics
agent/skills/audit-runtime-semantics - audit-strategy-idea
agent/skills/audit-strategy-idea - audit-strategy-spec
agent/skills/audit-strategy-spec - author-component
agent/skills/author-component - brainstorm-strategy-idea
agent/skills/brainstorm-strategy-idea - build-report-charts
agent/skills/build-report-charts - build-rule
agent/skills/build-rule - build-strategy-spec
agent/skills/build-strategy-spec - build-universe
agent/skills/build-universe - compare-experiments
agent/skills/compare-experiments - compare-strategy-versions
agent/skills/compare-strategy-versions - configure-trade-execution
agent/skills/configure-trade-execution - create-component
agent/skills/create-component - create-indicator
agent/skills/create-indicator - create-portfolio-optimizer
agent/skills/create-portfolio-optimizer - create-rule
agent/skills/create-rule - create-signal
agent/skills/create-signal - evaluate-cross-sectional
agent/skills/evaluate-cross-sectional - evaluate-factor
agent/skills/evaluate-factor - evaluate-time-series
agent/skills/evaluate-time-series - explore-data
agent/skills/explore-data - govern-research-workspace
agent/skills/govern-research-workspace - manage-live-trading
agent/skills/manage-live-trading - manage-strategy-version
agent/skills/manage-strategy-version - monitor-strategy-run
agent/skills/monitor-strategy-run - open-xquant
agent/skills/open-xquant - plot-indicators
agent/skills/plot-indicators - review-performance
agent/skills/review-performance - review-research-report
agent/skills/review-research-report - run-authorized-backtest
agent/skills/run-authorized-backtest - screen-factors
agent/skills/screen-factors - select-final-version
agent/skills/select-final-version - tune-parameters
agent/skills/tune-parameters - write-research-report
agent/skills/write-research-report
How it works
audit-artifact-lineage
When your agent uses it
Use when open-xquant artifacts must be checked for version/run/final traceability before comparison, report review, migration, or final selection.
What it loads into context
Frontmatter
43 tokensAlways available to the agent
SKILL.md
4,247 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
- Phase Path Containment Preflight
- Version Path Resolution
- Workspace Path Resolution
- Inputs
- Checks
- Transitive Current-Evidence Validation
- Lineage Audit Schema
- Selection-Bound Comparison Validation
- Outputs
- Red Lines
- Result
- Round 25 Revision Lineage
Files
SKILL.mdcore19,213 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubaudit-runtime-semantics
When your agent uses it
Compile an open-xquant strategy spec preview and audit that material SPEC execution semantics are preserved in compiled_plan.json before backtests.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
3,826 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
- Phase Path Containment Preflight
- Version Path Resolution
- Inputs
- Version-Governed Runtime Gate
- Confirmed Spec Audit Gate
- Runner Resolution
- Compile Preview
- Material Runtime Fields
- `runtime_audit.json`
- Output
Files
SKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytesSKILL.mdcore16,749 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubaudit-strategy-idea
When your agent uses it
Audit strategy_idea_brief.json and the brainstorm conversation before strategy_spec.yaml construction begins.
What it loads into context
Frontmatter
33 tokensAlways available to the agent
SKILL.md
2,298 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
- Phase Path Containment Preflight
- Version Path Resolution
- Inputs
- Version-Governed Artifact Gate
- Scope
- Strategy Idea Workflow Audit
- Default Confirmation Rule
- Output Artifact
- Result
Files
SKILL.mdcore10,123 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubaudit-strategy-spec
When your agent uses it
Audit strategy_spec.yaml field provenance before backtests.
What it loads into context
Frontmatter
19 tokensAlways available to the agent
SKILL.md
3,562 tokensWhen the skill is activated
Reference documents
7,776 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Phase Path Containment Preflight
- Version Path Resolution
- Reference Loading
- Inputs
- Version-Governed Artifact Gate
- Read-Only SPEC Boundary
- Runner Resolution
- Audited Idea Gate
- Spec Calibration Audit
- Provenance And Data Boundaries
- Component, Mapping, And Cross-Sectional Audit
- Gate
- Two-Step Spec Audit Gate
- Output
Files
SKILL.mdcore16,464 bytesSKILL.mdcore16,464 bytesSKILL.mdcore16,464 bytesSKILL.mdcore16,464 bytesSKILL.mdcore16,464 bytesreferences/component-and-mapping-audit.mddoc7,797 bytesreferences/confirmation-and-output.mddoc17,957 bytesreferences/field-classification.mddoc9,305 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubauthor-component
When your agent uses it
Author workspace-local open-xquant custom components from component_request.json, with tests, manifest hashing, catalog refresh, and no global SDK mutation.
What it loads into context
Frontmatter
39 tokensAlways available to the agent
SKILL.md
3,454 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
- Phase Path Containment Preflight
- Version Path Resolution
- Inputs
- Version-Governed Output Gate
- Output Layout
- Generated Cache Cleanup
- Workflow
- Cross-Sectional Component Feasibility
- Component Requirements
- Tests
- Manifest
- Result Artifact
- Red Lines
Files
SKILL.mdcore15,677 bytesSKILL.mdcore15,677 bytesSKILL.mdcore15,677 bytesSKILL.mdcore15,677 bytesSKILL.mdcore15,677 bytesSKILL.mdcore15,677 bytesSKILL.mdcore15,677 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubbrainstorm-strategy-idea
When your agent uses it
Guide a user through the open-xquant pre-spec strategy idea workflow and produce strategy_idea_brief.json before any strategy_spec.yaml work begins.
What it loads into context
Frontmatter
44 tokensAlways available to the agent
SKILL.md
2,559 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
- Phase Path Containment Preflight
- Version Path Resolution
- Scope
- Version-Governed Output Gate
- Conversation Rule
- Required Phases
- Brief Artifact
- Red Lines
- Result
Files
SKILL.mdcore11,317 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubbuild-report-charts
When your agent uses it
Use when report chart assets, figures, visual evidence, plotting scripts, or notebook-like assets are required for an open-xquant experiment report.
What it loads into context
Frontmatter
42 tokensAlways available to the agent
SKILL.md
6,387 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
- Phase Path Containment Preflight
- Version Path Resolution
- Version-Governed Chart Package
- Workflow
- Common Charts
- Canonical Report Chart Order
- Default Professional Chart Pack
- Chart Applicability Matrix
- Red Lines
- Current Immutable Report-Revision Contract
Files
SKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytesSKILL.mdcore27,407 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubbuild-rule
When your agent uses it
Configure and reason about open-xquant trading rules, exits, risk holds, and rebalance limits; use when the user asks for stop loss, take profit, drawdown guards, holding limits, or rule components.
What it loads into context
Frontmatter
57 tokensAlways available to the agent
SKILL.md
1,050 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
- Current Execution Model
- Built-in Rules
- Safe CLI Path
- SDK Pattern
- When A New Rule Is Needed
- Red Lines
Files
SKILL.mdcore4,280 bytesSKILL.mdcore4,280 bytesSKILL.mdcore4,280 bytes
Read SKILL.md
# Rule Builder
You help the user add risk and exit logic without overstating current spec
support.
## Current Execution Model
Rules are Python objects passed to `Engine.run(..., rules=[...])` or added by
the spec compiler internally. The audited CLI spec path does not support a
generic top-level `rules:` YAML section or `oxq backtest run --rules`, but it
does support a narrow whitelist under `portfolio.rules`.
In current `Engine.step()`:
View the complete source on GitHubbuild-strategy-spec
When your agent uses it
Build open-xquant strategy_spec.yaml files from audited strategy idea briefs for multi-Agent systems; stops after deterministic validation and writes a builder phase result for downstream orchestration.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
3,647 tokensWhen the skill is activated
Reference documents
6,366 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Phase Path Containment Preflight
- Version Path Resolution
- Read-Only Validation-Only Mode
- Reference Loading
- Version-Governed Artifact Gate
- Scope
- Audited Idea Input Gate
- Runner Resolution
- Provenance And SPEC Shape
- Data Inspection Boundary
- SPEC Audit Repair Handoff
- Build Flow
- Validate And Output
- Red Lines
Files
SKILL.mdcore16,482 bytesSKILL.mdcore16,482 bytesSKILL.mdcore16,482 bytesSKILL.mdcore16,482 bytesSKILL.mdcore16,482 bytesSKILL.mdcore16,482 bytesSKILL.mdcore16,482 bytesSKILL.mdcore16,482 bytesreferences/builder-output-schema.mddoc3,727 bytesreferences/source-mapping-and-catalog.mddoc14,442 bytesreferences/spec-shape-and-candidates.mddoc9,677 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubbuild-universe
When your agent uses it
Define open-xquant strategy universes and explain survivorship/PIT constraints; use when users choose symbols, indexes, dynamic filters, or tradable pools.
What it loads into context
Frontmatter
46 tokensAlways available to the agent
SKILL.md
618 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
- Current Stable Spec Path
- Required Checks
- SDK Path
- Red Lines
Files
SKILL.mdcore2,564 bytesSKILL.mdcore2,564 bytesSKILL.mdcore2,564 bytesSKILL.mdcore2,564 bytes
Read SKILL.md
# Universe Builder
You help the user define the tradable symbol pool.
## Current Stable Spec Path
The audited CLI compiler currently supports:
```yaml
universe:
type: static
symbols: ["SPY", "QQQ"]
point_in_time: false
survivorship_bias_policy: warnView the complete source on GitHubcompare-experiments
When your agent uses it
Compare two completed open-xquant experiment runs.
What it loads into context
Frontmatter
18 tokensAlways available to the agent
SKILL.md
913 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
- Preconditions
- Workflow
- Output Layout
- Report Evidence Gate
- Red Lines
Files
SKILL.mdcore4,205 bytes
Read SKILL.md
# Experiment Comparator
Use this skill when the user asks to compare two completed runs. The comparison
is cross-experiment metadata, so do not write outputs inside either run
directory. Resolve the comparison output root from `.open-xquant/workspace.yaml`
when it exists: use `paths.comparisons_dir` for the comparison directory root
and `paths.comparison_registry` for the summary registry. Fall back to
`comparisons/` and `comparisons/comparisons.jsonl` only when no workspace config
value is present.
Do not use this skill in a version-governed workspace. Stop before reading run
artifacts or writing comparison outputs and route the request to
`compare-strategy-versions`, including when the user only says "compare two
experiments". The version-aware comparator must validate each candidate againstView the complete source on GitHubcompare-strategy-versions
When your agent uses it
Use when comparing open-xquant runs or strategy versions, especially when distinguishing within-version reproducibility from cross-version strategy evidence.
What it loads into context
Frontmatter
41 tokensAlways available to the agent
SKILL.md
8,215 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
- Phase Path Containment Preflight
- Workspace Path Resolution
- Candidate Run Containment
- Final-Selection Comparison Handoff
- Immutable Selection Comparison Output
- Comparison Types
- Inputs
- Workflow
- Outputs
- Red Lines
- Result
- Round 25 Comparison Contract
Files
SKILL.mdcore33,602 bytesSKILL.mdcore33,602 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubconfigure-trade-execution
When your agent uses it
Configure open-xquant execution timing, fees, slippage, lot size, fill price mode, and broker assumptions; use when users discuss trading costs or order execution.
What it loads into context
Frontmatter
50 tokensAlways available to the agent
SKILL.md
1,135 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
- Stable Spec Configuration
- Cost Defaults Are Not Conclusions
- SDK Broker Pattern
- Red Lines
Files
SKILL.mdcore4,471 bytesSKILL.mdcore4,471 bytes
Read SKILL.md
# Trade Executor
You configure execution assumptions in `strategy_spec.yaml` or SDK brokers.
## Stable Spec Configuration
Use this for audited CLI backtests:
```yaml
signal:
signal_time: close_t
execution:
trade_time: next_openView the complete source on GitHubcreate-component
When your agent uses it
Route open-xquant component creation requests after checking the registry; use when users ask for a new Indicator, Signal, Rule, or PortfolioOptimizer.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
642 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
- Step 1: Classify The Request
- Step 2: Check Existing Registry
- Step 3: Route
- Red Lines
Files
SKILL.mdcore2,879 bytesSKILL.mdcore2,879 bytesSKILL.mdcore2,879 bytes
Read SKILL.md
# Component Creator
Use this skill when the user explicitly asks to modify open-xquant source code
or add a built-in component to the framework.
For workspace-local custom components in a research task, use
`author-component` instead. That worker writes `custom_components/`,
`component_manifest.json`, and a refreshed `component_catalog.json` without
mutating the installed SDK bundle.
When this built-in/framework development path applies, decide whether a new
component is actually needed, then route to the correct creation skill.
Do not write component code in this navigator skill.View the complete source on GitHubcreate-indicator
When your agent uses it
Create a new open-xquant Indicator with tests and registry wiring; use after create-component confirms no existing indicator satisfies the request.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
806 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
- Scope
- Phase 1: Read Existing Patterns
- Phase 2: Define Behavior
- Phase 3: Test First
- Phase 4: Implement
- Phase 5: Register
- Red Lines
Files
SKILL.mdcore3,376 bytesSKILL.mdcore3,376 bytesSKILL.mdcore3,376 bytesSKILL.mdcore3,376 bytes
Read SKILL.md
# Create Indicator
You create a pure numeric time-series component.
## Scope
Default built-in paths:
- source: `src/oxq/indicators/{snake_name}.py`
- tests: `tests/indicators/test_{snake_name}.py`
- package export: `src/oxq/indicators/__init__.py`
- built-in registry: `src/oxq/core/registry.py`
If the user is building a third-party extension instead of a built-in, useView the complete source on GitHubcreate-portfolio-optimizer
When your agent uses it
Create a new open-xquant PortfolioOptimizer with weight-invariant tests and registry wiring; use after create-component confirms no existing optimizer satisfies the allocation request.
What it loads into context
Frontmatter
47 tokensAlways available to the agent
SKILL.md
839 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
- Scope
- Phase 1: Read Existing Patterns
- Phase 2: Define Behavior
- Phase 3: Test First
- Phase 4: Implement
- Phase 5: Register
- Red Lines
Files
SKILL.mdcore3,425 bytesSKILL.mdcore3,425 bytesSKILL.mdcore3,425 bytes
Read SKILL.md
# Create PortfolioOptimizer
You create allocation logic that returns target weights.
## Scope
Default built-in paths:
- source: `src/oxq/portfolio/{snake_name}.py`
- tests: `tests/portfolio/test_{snake_name}.py`
- package export: `src/oxq/portfolio/__init__.py`
- built-in registry: `src/oxq/core/registry.py`
Existing built-ins live in `src/oxq/portfolio/optimizers.py`; read that fileView the complete source on GitHubcreate-rule
When your agent uses it
Create a new open-xquant Rule with portfolio-state tests and registry wiring; use after create-component confirms no existing rule satisfies the requested risk or exit behavior.
What it loads into context
Frontmatter
44 tokensAlways available to the agent
SKILL.md
854 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
- Scope
- Phase 1: Read Existing Patterns
- Phase 2: Define Behavior
- Phase 3: Test First
- Phase 4: Implement
- Phase 5: Register
- Red Lines
Files
SKILL.mdcore3,421 bytesSKILL.mdcore3,421 bytesSKILL.mdcore3,421 bytes
Read SKILL.md
# Create Rule
You create bar-by-bar risk, hold, weight override, or exit logic.
## Scope
Default built-in paths:
- source: `src/oxq/rules/{snake_name}.py`
- tests: `tests/rules/test_{snake_name}.py`
- package export: `src/oxq/rules/__init__.py`
- built-in registry: `src/oxq/core/registry.py`
## Phase 1: Read Existing PatternsView the complete source on GitHubcreate-signal
When your agent uses it
Create a new open-xquant Signal with deterministic output-domain tests and registry wiring; use after create-component confirms no existing signal satisfies the request.
What it loads into context
Frontmatter
43 tokensAlways available to the agent
SKILL.md
827 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
- Scope
- Phase 1: Read Existing Patterns
- Phase 2: Define Output Semantics
- Phase 3: Test First
- Phase 4: Implement
- Phase 5: Register
- Red Lines
Files
SKILL.mdcore3,409 bytesSKILL.mdcore3,409 bytesSKILL.mdcore3,409 bytes
Read SKILL.md
# Create Signal
You create a vectorized trading-intent component.
## Scope
Default built-in paths:
- source: `src/oxq/signals/{snake_name}.py`
- tests: `tests/signals/test_{snake_name}.py`
- package export: `src/oxq/signals/__init__.py`
- built-in registry: `src/oxq/core/registry.py`
## Phase 1: Read Existing PatternsView the complete source on GitHubevaluate-cross-sectional
When your agent uses it
Evaluate cross-sectional factors with IC, Rank IC, ICIR, decay, and turnover in open-xquant; use for stock selection and multi-asset ranking factors.
What it loads into context
Frontmatter
46 tokensAlways available to the agent
SKILL.md
615 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
- Minimal SDK Pattern
- Review Checklist
- Interpretation
- Report Shape
- Red Lines
Files
SKILL.mdcore2,448 bytesSKILL.mdcore2,448 bytes
Read SKILL.md
# Evaluate Cross-Sectional Factor
Use this when the factor ranks assets on the same date.
## Minimal SDK Pattern
```python
import pandas as pd
from oxq.data.market import LocalMarketDataProvider
from oxq.factor_eval.metrics import (
compute_decay,
compute_ic,
compute_icir,View the complete source on GitHubevaluate-factor
When your agent uses it
Route open-xquant factor evaluation tasks to cross-sectional or time-series workflows; use when users ask whether a factor predicts returns.
What it loads into context
Frontmatter
37 tokensAlways available to the agent
SKILL.md
333 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
- Ask First
- Route
- Data Requirements
- Red Lines
Files
SKILL.mdcore1,488 bytesSKILL.mdcore1,488 bytes
Read SKILL.md
# Factor Evaluator
You decide which factor evaluation workflow to use.
## Ask First
Confirm:
- factor definition
- symbols
- date range
- forward return horizons
- whether the question is stock selection or timing
- data source and missing-data treatmentView the complete source on GitHubevaluate-time-series
When your agent uses it
Evaluate time-series factors with hit rate, decay curve, profit/loss ratio, cash-period behavior, and tearsheets in open-xquant; use for timing or rotation signals.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
391 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
- Minimal SDK Pattern
- Review Checklist
- Interpretation
- Red Lines
Files
SKILL.mdcore1,668 bytesSKILL.mdcore1,668 bytesSKILL.mdcore1,668 bytes
Read SKILL.md
# Evaluate Time-Series Factor
Use this when the factor predicts direction through time for one asset or a
small rotation set.
## Minimal SDK Pattern
```python
from pathlib import Path
from oxq.factor_eval.bundle import create_bundle
from oxq.factor_eval.tearsheet import generate_tearsheet
factor_series = factor_df.stack().rename_axis(["date", "asset"]).rename("factor")View the complete source on GitHubexplore-data
When your agent uses it
Inspect, download, and validate open-xquant market, macro, and financial data; use before backtests, factor studies, or any task that depends on local parquet data.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
1,580 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
- Phase Path Containment Preflight
- Version Path Resolution
- Version-Governed Output Gate
- Runner Python
- Market Data
- Required Data Shape
- Macro Data
- Financial Data
- Red Lines
Files
SKILL.mdcore6,693 bytesSKILL.mdcore6,693 bytesSKILL.mdcore6,693 bytesSKILL.mdcore6,693 bytesSKILL.mdcore6,693 bytesSKILL.mdcore6,693 bytesSKILL.mdcore6,693 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubgovern-research-workspace
When your agent uses it
Use when an open-xquant research workspace looks disorganized, has root-level phase artifacts, or must be checked before phase handoff, comparison, migration, cleanup, or final selection.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,682 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
- Phase Path Containment Preflight
- Workspace Path Resolution
- Historical Version Audit
- Inputs
- Checks
- Outputs
- Journaled Governance Publication
- Red Lines
- Result
Files
SKILL.mdcore16,378 bytesSKILL.mdcore16,378 bytesSKILL.mdcore16,378 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubmanage-live-trading
When your agent uses it
Connect open-xquant to Alpaca paper or live trading; use only when users explicitly ask for broker connectivity, account checks, live data, or order submission.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
505 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
- Preconditions
- Account Check
- Market Data
- Order Submission
- LiveBroker
- Red Lines
Files
SKILL.mdcore2,115 bytesSKILL.mdcore2,115 bytesSKILL.mdcore2,115 bytesSKILL.mdcore2,115 bytesSKILL.mdcore2,115 bytesSKILL.mdcore2,115 bytesSKILL.mdcore2,115 bytes
Read SKILL.md
# Live Trader
You handle broker connectivity with strict safety gates.
## Preconditions
Install live dependencies:
```bash
uv sync --extra live
```
Set credentials:
View the complete source on GitHubmanage-strategy-version
When your agent uses it
Use when an open-xquant research conversation starts, resumes, or changes strategy meaning and the Agent must decide whether to continue a phase, create a new strategy version, or append a run.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,149 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
- Phase Path Containment Preflight
- Version Path Resolution
- Inputs
- Decision Rules
- Outputs
- Journaled Governance Publication
- Red Lines
- Result
Files
SKILL.mdcore13,891 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubmonitor-strategy-run
When your agent uses it
Audit, monitor, and record open-xquant strategy runs; use after backtests to run reproducibility checks, research bias checks, robustness, reports, and experiment logs.
What it loads into context
Frontmatter
50 tokensAlways available to the agent
SKILL.md
3,212 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
- Phase Path Containment Preflight
- Version Path Resolution
- Version-Governed Run Package
- Required Run Directory
- Audit
- Robustness
- Integrity Refresh Ordering
- Report And Experiment Log
- SDK Monitoring
- Red Lines
Files
SKILL.mdcore14,327 bytesSKILL.mdcore14,327 bytesSKILL.mdcore14,327 bytesSKILL.mdcore14,327 bytesSKILL.mdcore14,327 bytesSKILL.mdcore14,327 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubopen-xquant
When your agent uses it
Use when the user asks for any open-xquant or quantitative research task, including strategy design, backtesting, factor evaluation, parameter tuning, audit, robustness, reports, chart assets, SDK/component development, broker/live trading, workspace setup, Agent installation, or when deciding which open-xquant skill applies before CLI, SDK, tools, or file writes.
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
6,498 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
- Router Contract
- Runner And Workspace
- Workspace Classifier
- Task Routing
- Historical Schema-2 Staged Routing
- Install And Upgrade Questions
- Common Sequences
- Red Lines
- Round 25 Governance Routing
Files
SKILL.mdcore28,415 bytesSKILL.mdcore28,415 bytesSKILL.mdcore28,415 bytesSKILL.mdcore28,415 bytesSKILL.mdcore28,415 bytesSKILL.mdcore28,415 bytes
Read SKILL.md
# open-xquant Router
Round 26 routing: decision schema `5`; candidate, policy, comparison, and
lineage schema `3`; schema-4 pointers are rejected. Use the exact closed schema-3
`build_selection_comparison` request and route historical refresh as `write ->
review -> lineage -> prepare new selection -> comparison -> resume` with a fresh
selection and candidate-set hash; the old selection must not be reused. The
request includes `comparison_population`.
This is the mandatory entry skill for open-xquant work. It routes the task to
the most specific open-xquant skill and sets only the minimum runner and
workspace context needed for that handoff.
## Router ContractView the complete source on GitHubplot-indicators
When your agent uses it
Render open-xquant run charts and indicator overlays; use when users ask to visualize price, indicators, signals, or chart artifacts.
What it loads into context
Frontmatter
39 tokensAlways available to the agent
SKILL.md
325 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
- Current Tool Signature
- Dependency
- What To Check
- Red Lines
Files
SKILL.mdcore1,408 bytesSKILL.mdcore1,408 bytesSKILL.mdcore1,408 bytesSKILL.mdcore1,408 bytes
Read SKILL.md
# Chart Indicator
You create visual checks. Charts do not replace validation or audit.
## Current Tool Signature
`oxq.tools.chart.chart_indicator` plots indicator columns from a stored
`RunResult` in the tool session:
```python
from oxq.tools.chart import chart_indicator
result = chart_indicator(
run_id="run_1",View the complete source on GitHubreview-performance
When your agent uses it
Review open-xquant backtest performance, audit findings, reports, and experiment comparisons; use when users ask whether a strategy worked or how to interpret results.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,411 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
- Phase Path Containment Preflight
- Version Path Resolution
- Read Artifacts First
- Review Order
- Decision Language
- Compare Runs
- Red Lines
Files
SKILL.mdcore6,328 bytesSKILL.mdcore6,328 bytesSKILL.mdcore6,328 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubreview-research-report
When your agent uses it
Use when reviewing a completed open-xquant research report for decision consistency, artifact fidelity, audit/robustness interpretation, chart narrative quality, and semantic issues that deterministic report QA should not judge.
What it loads into context
Frontmatter
57 tokensAlways available to the agent
SKILL.md
6,685 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
- Phase Path Containment Preflight
- Version Path Resolution
- Version-Governed Review Gate
- Inputs
- Review Artifact Identity
- Governed Publication Lock
- Review Checklist
- Historical Schema-1 Output
- Result Invariants
- Red Lines
- Current Immutable Review Revision
Files
SKILL.mdcore29,118 bytesSKILL.mdcore29,118 bytesSKILL.mdcore29,118 bytesSKILL.mdcore29,118 bytesSKILL.mdcore29,118 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubrun-authorized-backtest
When your agent uses it
Run an authorized open-xquant backtest from gated artifacts without editing strategy specs, audits, runtime audits, or reports.
What it loads into context
Frontmatter
39 tokensAlways available to the agent
SKILL.md
3,167 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
- Phase Path Containment Preflight
- Version Path Resolution
- Inputs
- Version-Governed Run Gate
- Authorization Gate
- Runner Resolution
- Run
- Output
- Red Lines
Files
SKILL.mdcore13,572 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubscreen-factors
When your agent uses it
Screen symbols with open-xquant price, financial, and custom factors; use when users ask for value, quality, momentum, or multi-factor candidate lists.
What it loads into context
Frontmatter
44 tokensAlways available to the agent
SKILL.md
513 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
- Confirm Inputs
- Inspect Available Indicators
- Price-Based Screening Pattern
- Financial Screening
- Red Lines
Files
SKILL.mdcore2,066 bytesSKILL.mdcore2,066 bytesSKILL.mdcore2,066 bytes
Read SKILL.md
# Factor Screening
You build candidate lists from data. Screening is not a backtest.
## Confirm Inputs
Ask for:
- market and symbols
- factor definitions
- rebalance date or date range
- thresholds or ranking rules
- required data provider
- how to handle missing valuesView the complete source on GitHubselect-final-version
When your agent uses it
Use when a user asks to choose, confirm, promote, or mark a final open-xquant strategy version after audited runs, reports, and comparisons exist.
What it loads into context
Frontmatter
46 tokensAlways available to the agent
SKILL.md
16,639 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
- Workspace Path Resolution
- Phase Path Containment Preflight
- Candidate Resolution And Review Integrity
- Inputs
- Historical Schema-2 Policy Gate
- Eligibility Gate
- Staged Selection Orchestration
- Comparison Reference Ledger Publication
- Immutable Selection Comparison Output
- Candidate Set Schema And Publication
- Current Lineage Audit Gate
- Outputs
- Final Decision Schema And Publication
- Workspace-Scoped Selection Publication Lock
- Migration
- Red Lines
- Result
- Round 25 Current Selection Contract
Files
SKILL.mdcore70,840 bytesSKILL.mdcore70,840 bytesSKILL.mdcore70,840 bytesSKILL.mdcore70,840 bytesSKILL.mdcore70,840 bytesSKILL.mdcore70,840 bytes
Read SKILL.md
# Final Version Selector
Round 26: current decision schema `5` includes `report_revision` and
`selection_request_id`; schema-4 pointers are historical and rejected.
Candidate, policy, comparison, and lineage schema `3`.
This skill performs final version governance. It chooses a final research
candidate only from eligible audited candidates and only after the user
confirms the selection policy. It is not investment advice.
## Workspace Path Resolution
Read `.open-xquant/workspace.yaml` before reading or writing final-governance
artifacts. Resolve each key independently from `paths`: useView the complete source on GitHubtune-parameters
When your agent uses it
Tune open-xquant strategy parameters with grid search, walk-forward validation, time-series CV, and overfitting checks; use when users ask to optimize parameters.
What it loads into context
Frontmatter
47 tokensAlways available to the agent
SKILL.md
706 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
- Preconditions
- SDK Pattern
- Walk-Forward Required
- Time-Series CV
- Overfit Signals
- Red Lines
Files
SKILL.mdcore2,864 bytesSKILL.mdcore2,864 bytesSKILL.mdcore2,864 bytesSKILL.mdcore2,864 bytes
Read SKILL.md
# Parameter Tuner
You search parameters without turning in-sample luck into a claim.
## Preconditions
- The base strategy logic must already be clear.
- The base spec or SDK strategy must pass validation.
- Data must cover both training and OOS periods.
- The user must approve the metric and search ranges.
Warn when the total grid exceeds 100 combinations.
## SDK PatternView the complete source on GitHubwrite-research-report
When your agent uses it
Use when writing the final human-readable open-xquant experiment report from generated evidence, audit results, metrics, robustness output, and registered chart assets.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
5,583 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
- Phase Path Containment Preflight
- Version Path Resolution
- Mandatory Routing
- Version-Governed Report Package
- Language Parameter Gate
- Chart Decision Gate
- Inputs
- Writing Goal
- Institutional Report Standard
- Required Structure
- Fact Rules
- HTML Output
- Governed Publication Lock
- Red Lines
- Phase Result
- Current Immutable Candidate Publication
Files
SKILL.mdcore25,045 bytesSKILL.mdcore25,045 bytesSKILL.mdcore25,045 bytesSKILL.mdcore25,045 bytesSKILL.mdcore25,045 bytes
Read SKILL.md
## Phase Path Containment Preflight
Before any phase artifact read, write, directory creation, command, or handoff,
run this preflight completely and block on any failure:
1. Read `.open-xquant/workspace.yaml`. Resolve `version_root` from
`paths.versions_dir`, using `versions` only when that key is absent.
Require `version_root` to be a safe workspace-relative path: reject an
absolute path or any `..` path segment, resolve it canonically with
symlinks, and require the result to stay inside the workspace.
2. Read `current.json`. For normal phase work, set `expected_version_id` to
`current.json.active_version`; it must exist. Only a contract that
explicitly owns cross-version inspection may instead set
`expected_version_id` from the referenced version id for that historicalView the complete source on GitHubSkills included
- audit-artifact-lineage
- audit-runtime-semantics
- audit-strategy-idea
- audit-strategy-spec
- author-component
- brainstorm-strategy-idea
- build-report-charts
- build-rule
- build-strategy-spec
- build-universe
- compare-experiments
- compare-strategy-versions
- configure-trade-execution
- create-component
- create-indicator
- create-portfolio-optimizer
- create-rule
- create-signal
- evaluate-cross-sectional
- evaluate-factor
- evaluate-time-series
- explore-data
- govern-research-workspace
- manage-live-trading
- manage-strategy-version
- monitor-strategy-run
- open-xquant
- plot-indicators
- review-performance
- review-research-report
- run-authorized-backtest
- screen-factors
- select-final-version
- tune-parameters
- write-research-report
Repository facts
- License
- MIT
- Created
- Feb 26, 2026
- Last pushed
- Oct 12, 2026
- Forks
- 49
- Open issues
- 0
- 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 116 notes
agent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-runtime-semantics/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-strategy-spec/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-strategy-spec/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-strategy-spec/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/audit-strategy-spec/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/author-component/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/author-component/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/author-component/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/author-component/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/author-component/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/author-component/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/build-report-charts/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedagent/skills/build-report-charts/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- No broad tool permissions
- No dangerous shell commands
- 2 network referencesin 1 file: explore-data/SKILL.md:165
- No hidden or obfuscated text
- 1 instruction override signalin 1 file: build-rule/SKILL.md:27
- No executable scripts
- 2 references to secretsin 1 file: manage-live-trading/SKILL.md:24
- No requests to vote in Hot
Show all 5 signals by file
agent/skills/build-rule/SKILL.md
- Instruction override signalline 27
Do not tell the user that a `constraints` field will affect execution unless
agent/skills/explore-data/SKILL.md
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: 5 facts
- Instruction override signal in prose of the SKILL.mdbuild-rule/SKILL.md:27High confidence
- Network access in code of the SKILL.mdexplore-data/SKILL.md:165High confidence
- Network access in code of the SKILL.mdexplore-data/SKILL.md:166High confidence
- Access to secrets in code of the SKILL.mdmanage-live-trading/SKILL.md:24High confidence
- Access to secrets in code of the SKILL.mdmanage-live-trading/SKILL.md:25High confidence
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Path CLAUDE.md | .claude/skills/<skill-name>/ |
| Codex | Native | Manifest AGENTS.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md agent/skills/audit-artifact-lineage/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md agent/skills/audit-artifact-lineage/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md agent/skills/audit-artifact-lineage/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md agent/skills/audit-artifact-lineage/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.