本页还没有译文
本页 36 段还没译成Español,下面显示的是英文原文。intent stale
还没翻译译自英文(tanstack.com)本页 36 段都还没译
intent stale reports whether shipped skills may need review.
@tanstack/intent@latest stale [dir] [--json] [--github-review] [--package-label <label>]Options
--json: print JSON array of staleness reports--github-review: write review-reminder files and GitHub Actions output for the generated setup workflow--package-label <label>: set the fallback library label in generated reminder items
Behavior
Scope
- Checks the current package by default
- From a monorepo root, checks workspace packages that ship skills and also reports public workspace packages with no skill or artifact coverage
- Applies the
package.json#intent.skillsallowlist when falling back to installed dependencies; workspace packages are first-party and checked regardless. See Configuration. - When
diris provided, scopes the check to the targeted package or skills directory - Computes one staleness report per package
Coverage
- Reads repo-root
_artifacts/*domain_map.yamland_artifacts/*skill_tree.yamlwhen present - Flags public workspace packages that are not represented by generated skills or artifact coverage
- Skips workspace packages with
"private": true
Output and workflow state
- Prints text output by default or JSON with
--json - Text output lists flagged skills and review signals with their reasons; JSON contains only report data
- Prints a non-failing workflow update reminder when
.github/workflows/check-skills.ymlis missing the currentintent-workflow-versionstamp - If no packages are found, prints
No intent-enabled packages found.
Artifact coverage ignores can be recorded in _artifacts/*skill_tree.yaml or _artifacts/*domain_map.yaml:
coverage:
ignored_packages:
- '@tanstack/internal-tooling'
- name: packages/devtools-fixture
reason: test fixture onlyIgnored packages are excluded from missing coverage signals. Private workspace packages are excluded automatically.
JSON report schema
--json outputs an array of reports:
[
{
"library": "string",
"currentVersion": "string | null",
"skillVersion": "string | null",
"versionDrift": "major | minor | patch | null",
"skills": [
{
"name": "string",
"reasons": ["string"],
"needsReview": true
}
],
"signals": [
{
"type": "missing-package-coverage",
"library": "string",
"subject": "string",
"reasons": ["string"],
"needsReview": true,
"packageName": "string",
"packageRoot": "string"
}
]
}
]Report fields:
| Field | Meaning |
|---|---|
library | Package name |
currentVersion | Local package version when available; otherwise latest from npm, or null if unavailable |
skillVersion | library_version from skills, or null |
versionDrift | major, minor, patch, or null |
skills | Per-skill checks |
signals | Artifact and workspace coverage checks |
Skill fields:
namereasons: one or more staleness reasonsneedsReview: boolean (truewhen reasons exist)
Reason generation:
version drift (<skillVersion> → <currentVersion>)new source (<path>)when a declared source has no stored sync SHA- artifact parse warnings, unresolved artifact skill paths, source drift, artifact library version drift, and missing workspace package coverage
Text output
- Report header format:
<library> (<skillVersion> → <currentVersion>) [<versionDrift> drift] - When no skill reasons exist:
All skills up-to-date - Otherwise: one warning line per stale skill or review signal (
⚠ <name>: <reason1>, <reason2>, ...) - Review signals that produce the same line, such as one raised by both
domain_map.yamlandskill_tree.yaml, print once; JSON output keeps each signal with itsartifactPath
Generated review files
The existing --github-review workflow mode writes review-items.json and, when there are items, pr-body.md. Its Agent Review instructions route to generate-skill and its conditional review-signals reference. The command itself does not create a remote PR or edit skills; the installed GitHub workflow handles the review reminder.
Version drift and missing stored source SHAs do not prove that guidance is wrong. Review actual source changes before editing. Missing evidence stays unresolved, and an evidence-backed no-op may leave a signal flagged. Failed checks require logs; workflow advisories are separate maintenance items. Neither is a reason to rewrite skills.
stale and review serve different checks
| Command | Evidence | Persistent result |
|---|---|---|
intent stale | Installed or registry package versions, stored source sync SHAs, artifacts, workflow version, and workspace package coverage. | Reports conservative signals; it does not record semantic review. |
intent review | Git changes plus recorded source, skill, and planning-record fingerprints. | Records evidence-backed outcomes in .intent/review-state.json and reopens them when tracked content changes. |
Common errors
- Package scan failure: prints a scanner error
- Registry fetch failures do not crash command;
currentVersionmay benull
Notes
- Source staleness checking is conservative: it flags missing source SHAs in sync-state, not remote content differences.