xingwudao/open-xquant/audit-runtime-semantics

Claude Code and agent skill by xingwudao/open-xquant

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Compile an open-xquant strategy spec preview and audit that material SPEC execution semantics are preserved in compiled_plan.json before backtests.

Install

npx vskill@latest install xingwudao/open-xquant/audit-runtime-semantics

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.

  1. 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 historical read. This exception never permits active-version work to consume another version.

  1. Set the intended version directory to

<version_root>/<expected_version_id>/ and resolve it canonically. Require the intended version directory to remain inside the canonical version root and workspace; otherwise treat it as a symlink escape. Read version_manifest.json only from that exact directory. The manifest version_id must equal expected_version_id; for normal phase work it therefore must also equal current.json.active_version.

  1. Before using each required phase_paths value, require a non-empty

workspace-relative string. Reject an absolute path and any .. path segment. Resolve <workspace>/<phase_path> canonically, including existing symlink ancestors even when the leaf will be created, and require the target to be the intended version directory or a descendant of it. A symlink escape outside that directory is invalid.

  1. On any identity or path failure, stop before phase artifact reads, writes,

directory creation, commands, or handoffs. Do not normalize an unsafe path into acceptance and do not fall back to a default phase path.

Block examples when expected_version_id is v001 include strategy_store/v001/../v002/04_spec_build, strategy_store/v002/04_spec_build, /tmp/04_spec_build, and strategy_store/v001/escape/04_spec_build when escape is a symlink whose target is outside the intended version directory. An allowed custom nested phase path is strategy_store/v001/custom/phases/04_spec_build when its canonical target remains under the intended version directory.

For a new-version bootstrap, only manage-strategy-version may proceed before the new manifest exists or the new id becomes active. It must apply the same workspace-relative, traversal, canonical-containment, and symlink checks to every constructed phase path before directory creation, then write a matching manifest before publishing current.json last.

Version Path Resolution

Before using any <phase_paths.*> or <version_root> placeholder, read .open-xquant/workspace.yaml. Resolve version_root from paths.versions_dir; use versions only when that key is absent. Require a safe relative path whose resolved target stays inside the workspace. Then read <version_root>/<version_id>/version_manifest.json and use its exact phase_paths entry for each phase. For example, a configured root of research_versions must resolve the spec-build phase to research_versions/v003/04_spec_build; never redirect it to a default-root phase path.

Resolve components_dir independently from paths.components_dir; use components only when that key is absent. Require a safe relative path whose resolved target stays inside the workspace. Use the resolved <components_dir> for every workspace-local component bundle path.

Resolve conversations_dir independently from paths.conversations_dir; use conversations only when that key is absent. Require a safe relative path whose resolved target stays inside the workspace. Reject absolute paths, traversal outside the workspace, and symlink escapes whose resolved target leaves the workspace. Use the resolved <conversations_dir> for every confirmation-event read.

Runtime Auditor

Use this skill after a confirmed spec_audit.json from audit-strategy-spec and before any formal oxq backtest run. Its job is deterministic artifact consistency: strategy_spec.yaml must compile into a compiled_plan.json that preserves material execution semantics.

This skill does not audit conversation provenance, ask the user to confirm assumptions, edit strategy_spec.yaml, run a backtest, monitor a run, or write a report.

Inputs

  • strategy_spec.yaml
  • spec_audit.json
  • component_catalog.json when available
  • backtest_authorization.json or the intended formal run inputs when

available, including data_dir and component_manifest paths

  • Optional prior <phase_paths.07_compile_preview>/compiled_plan.json

Version-Governed Runtime Gate

Before compiling, read current.json and use active_version as version_id. If no active version exists, block and return to manage-strategy-version.

Read the confirmed SPEC package from:

text
<phase_paths.04_spec_build>/strategy_spec.yaml
<phase_paths.06_spec_audit>/spec_audit.json
<phase_paths.04_spec_build>/component_catalog.json

Write compile preview artifacts only under:

text
<phase_paths.07_compile_preview>/compiled_plan.json
<phase_paths.07_compile_preview>/strategy.py
<phase_paths.07_compile_preview>/spec_hash.txt

Write the runtime audit only to:

text
<phase_paths.08_runtime_audit>/runtime_audit.json

Do not write root-level compile_preview/, compiled_plan.json, or runtime_audit.json.

Confirmed Spec Audit Gate

Before compiling, read spec_audit.json and block unless all of these are true:

  • status: pass
  • schema_version: 4
  • spec_provenance_pass: true
  • audit_conclusion: all_pass
  • user_confirmation_status: confirmed
  • confirmation_event exists, points to

<conversations_dir>/<conversation_id>/confirmations.jsonl, records the full SPEC table path/hash, and records spec_audit_path plus spec_audit_hash for the pre-confirmation spec_audit.json. The JSONL line must match the audit reference for path, event_id, decision: confirmed, line_number, artifact_path, artifact_hash, spec_audit_path, and spec_audit_hash. A rejected or missing machine-readable decision blocks strict confirmation.

confirmation_event.event_hash is the SHA-256 hash of the raw JSONL line content referenced by path and line_number. Do not compare it with a nested event_hash field inside the JSONL payload if one exists; that payload field is not part of the formal gate. Likewise, confirmation_event.spec_audit_hash is the canonical pre-confirmation audit hash, not the current final post-confirmation spec_audit.json hash. The deterministic spec-audit validate --strict-confirmed --json command is the source of truth for confirmation-event binding.

Before compiling, run deterministic validation and block unless it exits 0:

bash
<resolved_runner> spec-audit validate <phase_paths.06_spec_audit>/spec_audit.json \
--spec <phase_paths.04_spec_build>/strategy_spec.yaml \
--strict-confirmed \
--json

An audit_conclusion: all_pass audit with user_confirmation_status: pending is still blocked. Do not treat it as authorization for runtime audit, compile preview, or backtest work.

Runner Resolution

In a research workspace, do not assume uv run oxq is installed locally. Before running deterministic validation or compile commands:

  1. Read ~/.config/open-xquant/agent.yaml.
  2. Prefer preferred_runner_argv when the shell tool accepts argv; otherwise

use preferred_runner in place of uv run oxq or bare oxq.

  1. If it is missing or fails, read ~/.config/open-xquant/agent-install.json,

take sdk_bundle.runner.argv or sdk_bundle.runner.oxq, and use that cached runner.

Keep the shell in the user's research directory. Do not search unrelated home directories for another open-xquant checkout.

Compile Preview

If no current compile preview exists for the current spec_hash, run:

bash
<resolved_runner> strategy compile <phase_paths.04_spec_build>/strategy_spec.yaml \
--data-dir data \
--component-manifest <components_dir>/bundles/<bundle_id>/component_manifest.json \
--out <phase_paths.07_compile_preview>

Use the same data_dir and every component_manifest path that the formal oxq backtest run will use. Omit --data-dir only when the formal run will also omit it. Omit --component-manifest when no workspace-local custom components are authorized; repeat it for each authorized bundle manifest. The preview compiled_plan.json includes the resolved effective data_dir, so a preview made with different run inputs is not a valid runtime gate. When component manifests are used, record their authorized bundle_hash values in runtime_audit.json as component_bundle_hashes.

Read <phase_paths.07_compile_preview>/compiled_plan.json and <phase_paths.07_compile_preview>/spec_hash.txt. The spec_hash in the compile preview must match strategy_spec.yaml and spec_audit.json.

Read the compiled plan's open_xquant_version and compare it with strategy_spec.yaml field required_oxq_version. When a prior or preview runtime artifact includes environment.json, use it as additional evidence for the same version comparison. A compile preview produced by a different open-xquant package version than the SPEC requires is not a valid runtime gate.

The compile preview must also include <phase_paths.07_compile_preview>/strategy.py. Read the complete generated file and compute strategy_source_hash as the full SHA-256 of its raw bytes. Record the path and hash in runtime_audit.json and return the path, hash, and complete source text to the coordinator. Use a handoff field named strategy_source_code; a path-only handoff is not enough.

The runtime auditor owns generated-source identity, not user-facing presentation. It must not write source-presentation evidence, set a presentation boolean, or claim that the coordinator relayed the source. The coordinator displays the exact strategy_source_code in a fenced python block to print the complete strategy.py source and records durable presentation evidence only after that relay.

Material Runtime Fields

Compare material SPEC fields against compiled_plan.json. At minimum audit:

  • rebalance interval and runtime rebalance source
  • calendar rebalance frequency and schedule
  • execution timing and fill price mode
  • fee, side-specific fees, sell-side tax, minimum fee, and slippage
  • universe type, index metadata, point-in-time policy, and constituent symbols
  • data filters, suspension_policy, and required filter columns
  • indicator component classes, params, and lag_bars
  • cross-sectional indicator preservation, including RPS when present
  • portfolio optimizer params that affect allocation, including

TopNRanking.pre_filter_signal, weighting, and ascending

  • validation train/test periods and required_oos
  • rule components that affect orders or positions
  • portfolio constraints that affect target weights
  • initial cash, cash return, benchmark, and metrics profile when represented in

runtime artifacts

  • required_oxq_version against the open-xquant open_xquant_version recorded

by compile/runtime artifacts such as compiled_plan.json and environment.json when available

If the compiled plan omits a supported material field, or contradicts the SPEC, block the backtest. For example, if strategy_spec.yaml says portfolio.rules.rebalance.params.interval_days: 10 but compiled_plan.json shows daily or interval_days: 1, this is a blocking runtime mismatch.

If the engine cannot represent a material field in compiled_plan.json, block and report that the runtime artifact is insufficient for formal execution. Do not infer success from a missing field.

This is the start of the file. Read the full SKILL.md on GitHub.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/xingwudao/open-xquant/audit-runtime-semantics/badge)](https://verified-skill.com/skills/xingwudao/open-xquant/audit-runtime-semantics)

Something wrong with this skill or its page? Tell us.