CtrlK
BlogDocsLog inGet started
Tessl Logo

fix-local-checks

Patch the code so `scripts/local-checks.sh` passes, after /implement-mainspec — a narrow post-implement polish specialist. Reads the (remediation-rich) check failures, fixes the underlying cause, never silences a check, re-verifies, and commits. Invoked headless by the dispatcher's two-strike local-checks gate; the dispatcher owns the retry counter. Triggers - fix-local-checks, fix lint, fix local checks, fix typecheck, make checks pass, post-implement polish (project)

68

Quality

85%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

The canonical home for this skill is fix-local-checks in tdg-ninja/context-specs-factory-ai

SKILL.md
Quality
Evals
Security

Quality

Content

73%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A tightly scoped, well-sequenced specialist skill with explicit verification and stop paths, but it pays a token tax by restating the same prohibitions in three separate sections. Moving reference material out of SKILL.md and deduplicating the never-lists would lift it further.

Suggestions

Consolidate "The one thing that matters", "You may NEVER silence a check", and "Hard nevers" into one section — the three lists overlap almost completely and cost ~20 lines for no added constraint.

Move the exhaustive suppression-directive enumeration and the dispatcher-contract details into a `references/` file, keeping SKILL.md as a lean overview with one-level-deep pointers.

Add one brief worked triage example (a real failure message → the honest fix it points at) to sharpen the needs-judgment branch, which is currently the only vague direction.

DimensionReasoningScore

Conciseness

Mostly lean and imperative with no concept re-teaching, but the core prohibitions are stated three times: "**Every one of those is forbidden**" in the opening section, the full "You may NEVER silence a check" list, and then "Hard nevers" which restates "Never silence a check", "Never add a test-skip marker", and "Never touch the check machinery" already covered above. Roughly 15–20 lines could be cut without losing a single constraint. Not 2 because there is no padding or explanation of things Claude already knows; not 4 because the duplication is more than a minor trim — it is a structural restatement of the same list.

3 / 5

Actionability

Concrete commands throughout: `./scripts/local-checks.sh`, `./prds/<f>/run-prd-test.sh`, `claude -p "/fix-local-checks <feature>"`, a model commit message (`fix(local-checks): trim query whitespace`), and an exact enumeration of forbidden directives (eslint-disable*, @ts-ignore, # noqa, #[allow(...)]). Not 5 because some directions stay judgment-shaped ("Do the best **honest** fix you can", "Make whatever honest partial progress you can") without a worked example of triage in action; not 3 because everything runnable is spelled out exactly.

4 / 5

Workflow Clarity

Steps 1–5 are a clear, ordered sequence with an explicit validation checkpoint ("Re-run `./scripts/local-checks.sh`" before commit), a feedback loop (dispatcher re-runs the gate; partial-progress commits so attempts compound), and a defined stop/error-recovery path (step 5, STUCK escalation). Not 4 because validation is explicit at the commit boundary and both success and failure outcomes are fully specified.

5 / 5

Progressive Disclosure

Well-organized single-file skill with clear section headers and no bundle files, so nothing is mis-filed or nested; navigation is easy. Not 5 because at ~105 lines it exceeds the simple-skill (<50 lines) exception and some reference-shaped material (the exhaustive suppression-directive list, the dispatcher contract) could live in a references file with SKILL.md as overview; not 3 because the structure that is present is clean and nothing that belongs elsewhere is buried.

4 / 5

Total

16

/

20

Passed

Description

92%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A highly effective description: concrete third-person actions, explicit both-ways trigger guidance, and strong niche scoping tied to a specific script and pipeline position. Only gap is a few natural trigger synonyms.

Suggestions

Add a few natural trigger variations users might actually say, e.g. "make lint pass", "fix tests", "make CI green", to broaden keyword coverage.

"Invoked headless by the dispatcher's two-strike local-checks gate" assumes pipeline context a fresh reader lacks; a brief plain-language clause on when a human should reach for it would help.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — "Reads the (remediation-rich) check failures, fixes the underlying cause, never silences a check, re-verifies, and commits" — comprehensively covering the skill's job in third-person imperative voice. Not level 4 because coverage of the what is complete rather than having minor gaps; not below because there is no vague or generic language.

5 / 5

Completeness

Explicitly answers what ("Patch the code so `scripts/local-checks.sh` passes", "fixes the underlying cause, never silences a check, re-verifies, and commits") and when ("after /implement-mainspec", "Invoked headless by the dispatcher's two-strike local-checks gate", plus an explicit "Triggers -" clause). Not 4 because the when is fully explicit with concrete trigger phrases rather than merely present.

5 / 5

Trigger Term Quality

"Triggers - fix-local-checks, fix lint, fix local checks, fix typecheck, make checks pass, post-implement polish" gives good natural-phrase coverage. Not 5 because common variations users would say are missing (e.g. "make lint pass", "fix the build", "fix tests", "make CI green"); not 3 because multiple natural synonyms are present, not just one or two generic keywords.

4 / 5

Distinctiveness Conflict Risk

"a narrow post-implement polish specialist" scoped to a named script (`scripts/local-checks.sh`), a named predecessor skill (/implement-mainspec), and the dispatcher's gate carves a clear niche with distinct triggers and minimal conflict risk. Not 4 because the unambiguous "fix-local-checks" / "fix local checks" anchors the niche beyond the mildly generic "fix lint" / "make checks pass" phrases.

5 / 5

Total

19

/

20

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

referenced_paths_exist

Referenced path issues: 4 missing

Warning

Total

15

/

16

Passed

Repository
tdg-ninja/context-specs-claude-code
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.