CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

qa-bug-fixer

Receive bug reports from QA agents and implement focused, minimal fixes. Read before writing, understand the root cause, fix with the smallest possible change, and report what was changed. Works with any tech stack. Trigger on "fix the QA bugs", "implement the bug fix", "patch the defect", or when a QA agent has produced bug reports with reproduction steps.

Quality

80%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./.claude-plugin/plugins/qa-bug-fixer/skills/qa-bug-fixer/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

63%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.

The body delivers a clear, actionable three-mode workflow with concrete tool names, an output template, and error-recovery guidance. Its two real weaknesses are redundancy (Persona and closing Workflow sections duplicate earlier content) and a disclosure scheme that references rules/*.md files missing from the bundle, so the mandated standards are unavailable to the reader.

Suggestions

Include the referenced rules/ files (std-minimal.md, std-read-first.md, std-report.md, _sections.md) in the bundle, or inline their content into SKILL.md, since 'Every fix must comply with rules in the rules/ directory' currently points at files that do not exist.

Delete the Persona section and the closing 'Workflow' section — both restate the opening paragraph and the Mode B steps — to cut roughly 20 lines of duplication.

Strengthen the Mode B Verify step with a concrete checkpoint (e.g., run the specific test class covering the changed file, or re-run the reproduction steps from the bug report) so batch Mode A fixes get real validation rather than a coverage check.

DimensionReasoningScore

Conciseness

The body is mostly efficient — tables for mode detection and rules, a compact output template — but the Persona section ("Role: Senior Engineer — surgical bug fixer", "Attitude: Minimal, reads before writing") and the closing "Workflow" section largely restate the opening paragraph and the Mode B steps. Removing those duplicated sections would tighten it without losing information, which fits the 'mostly efficient but could be tightened' anchor rather than anchor 4's minor-trim case.

3 / 5

Actionability

Guidance is largely executable: exact MCP tool names ("mcp__linear__save_comment", "mcp__github__add_issue_comment"), a concrete ".qa/config.yml" check, a copy-ready fix report template, and severity-ordered batching. It falls short of anchor 5 because "Every fix must comply with rules in the `rules/` directory" and "See `rules/_sections.md` for section definitions" point to files that are not present in the bundle, leaving referenced rule content unavailable.

4 / 5

Workflow Clarity

Mode B gives a clearly sequenced Read → Understand → Fix → Verify → Report → Update tracker workflow, and the Troubleshooting section provides error-recovery loops (missing reproduction steps → ask; scope beyond 3 files → flag for review). The Verify checkpoint is somewhat weak — "checking if existing tests cover the fix" rather than running them, and batch Mode A relies on delegated Mode B verification — so it sits at anchor 4 rather than the fully validated anchor 5.

4 / 5

Progressive Disclosure

The in-body structure is good (clear sections, a rules table with per-file impact levels, references clearly signaled one level deep), but all four referenced bundle files ("rules/std-minimal.md", "rules/std-read-first.md", "rules/std-report.md", "rules/_sections.md") are absent from the bundle — there is no rules/ directory at all. The disclosure scheme is well designed on paper but cannot actually be navigated, which is a better fit for anchor 3's 'could be better organized' than anchor 4's appropriately-placed content.

3 / 5

Total

14

/

20

Passed

Description

88%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 strong description: it names the domain and multiple concrete actions, uses third person, and includes explicit trigger phrases for both batch and single-fix requests. The 'when' guidance is explicit and natural, with only minor overlap risk against generic bug-fixing requests and a few missing trigger synonyms.

DimensionReasoningScore

Specificity

The description lists multiple concrete actions covering the whole workflow: "Receive bug reports from QA agents", "implement focused, minimal fixes", "understand the root cause", "fix with the smallest possible change", and "report what was changed". Coverage is comprehensive rather than gapped, matching the anchor for multiple specific concrete actions rather than the anchor-4 case of minor gaps.

5 / 5

Completeness

Both "what" ("Receive bug reports... and implement focused, minimal fixes... report what was changed") and "when" ("Trigger on 'fix the QA bugs'...") are explicitly present with concrete trigger phrases. However the "when" clause leans on quoted trigger strings plus a QA-agent-production condition rather than a plain 'Use when...' statement, leaving it slightly less explicit than the anchor-5 example.

4 / 5

Trigger Term Quality

Trigger phrases "fix the QA bugs", "implement the bug fix", "patch the defect", and "when a QA agent has produced bug reports with reproduction steps" are natural phrasings with good synonym coverage (fix/patch/defect/bug report). It stops short of anchor 5 because common variations like "resolve the bug", "repair", or "address the failing test" are absent.

4 / 5

Distinctiveness Conflict Risk

The QA-agent-report niche ("bug reports from QA agents", "a QA agent has produced bug reports") is mostly distinct from general coding skills. Minor overlap remains: "implement the bug fix" and "patch the defect" could plausibly fire on ordinary non-QA bug-fix requests, fitting the anchor-4 'minor overlap risk with closely related skills' case.

4 / 5

Total

17

/

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

metadata_field

'metadata' should map string keys to string values

Warning

Total

15

/

16

Passed

Repository

ravnhq/ai-toolkit

Path

.claude-plugin/plugins/qa-bug-fixer/skills/qa-bug-fixer/SKILL.md

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.