verify-skill-migration
Validator-backed migration QA for Codex skills and related docs
- License 许可
- MIT
- 最近更新
- 2026-08-20
- Stars
- ★ 75
- Forks
- 5
01 前言 Front Matter
- name
- verify-skill-migration
- description
- Validator-backed migration QA for Codex skills and related docs
02 资源文件 1
- 脚本scripts/validate_skill_migration.py39.9 KB
03 SKILL.md 正文
<!-- Generated by the ccode-to-codex mirror sync on 2026-04-14 from input revision 9a2403fb. -->
Verify Skill Migration
Use the canonical migration validator after each batch and before finalizing a migrated skill set.
The validator implementation lives at:
.codex/skills/verify-skill-migration/scripts/validate_skill_migration.py
Use the stable repo entrypoint for normal runs:
python3 tools/migration_support/validate_names.py
Checks
structural-integrity
- Fails on MCP callable namespaces that still preserve hyphenated server IDs
frontmatter
- Applies to
SKILL.mdentrypoints only - Parses YAML frontmatter
- Requires
nameanddescription - Fails on prohibited keys:
allowed-tools,effort,metadata
line-count
- Warns when a
SKILL.mdfile reaches 500+ lines - Warning only, not a failure
references
- Validates
references/directories under the scan scope - Fails on empty
references/directories - Fails on empty reference files, including nested files
Cross-References Check
- Warns on unknown skill references
- Uses discovered skill names instead of treating arbitrary backticked words as skills
skill-remnants
- Fails on legacy skill primitive call syntax in Codex artifacts
stale-paths
- Fails on operational legacy migration path references
- Ignores provenance banner lines such as
Source:andGenerated by:
operator-guidance
- Applies to the migration-family skill packages
- Fails on operator-facing repo paths that do not resolve there
- Fails on mistaken filesystem-root repo paths such as
/.codex/... - Fails on legacy slash-skill invocations in migration-family operational guidance
- Resolves skill-local
references/...andscripts/...before repo-root paths
agent-dependencies
- Fails on unresolved named agent runtime dependencies
- Treats only
.codex/agents/*.tomlas authoritative runtime contracts - Does not let legacy
.codex/agents/*.mdreferences satisfy dependency resolution
agent-flattening
- Warns when
spawn_agent(...)appears to lose explicitopus/sonnet/haikuruntime intent - Looks for nearby model-intent context without matching
modelorreasoning_effortoverrides
agent-api-shape
- Warns on migrated-artifact examples that still use legacy positional
wait_agent(...)call shapes - Targets migration references and Tier 3 migrated skill artifacts without blocking unrelated older skills
executable-primitives
- Fails on Claude primitives in operational Codex guidance
- Ignores migration/comparison references and mapping-table contexts where primitive names are discussed, not executed
environment-readiness
- Inspects an explicit Codex home, defaulting to repo-local
.codex - Parses
config.tomlwhen present - Warns on missing MCP commands by default because PATH is host-dependent
- Parses optional
hooks.jsonand checks referenced hook commands/scripts - Warns on oversized or legacy-reference-heavy
AGENTS.md - Validates plugin manifest JSON when plugin files exist
- Use
--strict-environment-readinesswhen these host warnings should fail
Usage
Run the full validator during post-flight:
python3 tools/migration_support/validate_names.py
Run a targeted scope:
python3 tools/migration_support/validate_names.py \
--scan-dir .codex/skills/verify-skill-migration
Run specific checks only:
python3 tools/migration_support/validate_names.py \
--check frontmatter \
--check cross-references \
--check operator-guidance \
--check agent-dependencies \
--check agent-api-shape
Run environment readiness against the repo-local Codex config:
python3 tools/migration_support/validate_names.py \
--check environment-readiness \
--codex-home .codex
List available checks:
python3 tools/migration_support/validate_names.py --list-checks
Result Semantics
FAIL: migration blockers or invalid Codex artifactsWARN: quality issues that should be reviewed but do not block executionPASS: no findings
04 同仓库更多 Skills
hello-world-skill
Create or update hello-world.txt with a Codex greeting. Use when you need a minimal instruction-only Agent Skill example with explicit verification.
migration-dashboard
Codex-native migration analysis dashboard for Claude→Codex skills and agents. Reads `migration-state.json` to report progress, drift, velocity, nativeness-review state, validator results, and next-batch recommendations. Use for "migration status", "what should I migrate next", "migration progress", or "migration dashboard".
migrate-to-codex
Migrate Claude-era skills into native Codex skill artifacts with a preservation-first workflow. Also migrate project-local Claude slash commands and command hooks in batch mode as review-required Codex artifacts. Classify migration risk, preserve control shape, validate runtime-coupling leaks, and require a Codex nativeness audit before calling the result native. Use for "migrate skill to codex", "convert skill to codex", "port skills", or "codex compatibility". This skill does not migrate agents.
migrate-agents-to-codex
Migrate Claude agent definitions (.claude/agents/*.md) to Codex-compatible markdown or TOML outputs under `.codex/agents/`. Strips frontmatter, maps fields to native Codex contracts, normalizes MCP namespaces, and classifies agents into 3 tiers. Includes adversarial cross-model review (GPT-5.4 inline native sub-agent) with auto-fix loop before completion. Use when asked to "migrate agents to codex", "convert agents", or "port agent definitions". Pipeline stage: CODE → VERIFY.
example-skill
Minimal placeholder skill for public migration examples.
example-commit-helper
Minimal placeholder commit helper for public migration examples.