skill-debug

Debug a reproducible symptom with a bounded feedback loop and original-scenario verification

Quality

58%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/skill-debug/SKILL.md

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 is a lean, disciplined debugging procedure with unusually strong validation culture (hard gates, anti-rationalization checks, bounded retries). Its main weaknesses are that the central feedback-loop procedure is outsourced to a plugin file not present in this bundle, and several references point to environment-specific paths that cannot be verified from the skill itself.

Suggestions

Inline the essential feedback-loop steps from 'skills/blocks/debug-feedback-loop.md' (or ship it as a bundle reference file) so the core procedure is actionable without relying on plugin-internal paths.

Clarify or bundle the external dependencies — '~/.claude-octopus/loop-config.conf', '/octo:unfreeze', 'THIRD_PARTY_NOTICES.md' — either by including them in the skill directory or stating their expected provenance, so navigation does not dead-end.

Trim the WTF-score reporting block to the essential rule and threshold, moving the scoring defaults to a reference file if they are needed at all.

Dimension	Reasoning	Score
Conciseness	The body is dense and assumes competence ('Confidence is not verification', 'Do not treat a nearby passing helper test as proof') with no explanation of concepts Claude already knows; only the Octopus/WTF-score and freeze-guard bookkeeping runs slightly long, keeping it below anchor 5.	4 / 5
Actionability	Concrete guidance exists (3-Strike Rule with explicit thresholds, an executable bash freeze snippet, the 'inconclusive' return value, the WTF-score formula), but the core feedback-loop procedure is delegated to 'skills/blocks/debug-feedback-loop.md', which is not part of this bundle — the key executable detail is missing, matching anchor 3 rather than anchor 4's 'minor gaps'.	3 / 5
Workflow Clarity	A clear sequence is present (reproduce the symptom → minimize while retaining the failure signature → test one named hypothesis → verify both minimal repro and original scenario → remove instrumentation → keep a regression test) with abundant explicit checkpoints (the HARD-GATE, the anti-rationalization check, mandatory strategy rotation); only the delegated external loop file keeps it from anchor 5.	4 / 5
Progressive Disclosure	Sections are well-organized and the skill is short enough not to need bundle files, but every referenced path ('skills/blocks/engineering-method-selection.md', 'skills/blocks/debug-feedback-loop.md', '~/.claude-octopus/loop-config.conf', 'THIRD_PARTY_NOTICES.md') points outside the bundle to files that do not exist here — references present but not clearly signaled or resolvable.	3 / 5
	Total	14 / 20 Passed

Description

53%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.

The description communicates a clear debugging niche with a methodological identity, but it reads as a policy statement rather than an invocable trigger: it lacks any when-to-use guidance and misses the everyday vocabulary (bug, error, failing test) users actually say. Its distinctiveness is good but its trigger surface is thin.

Suggestions

Add an explicit 'Use when...' clause to the description itself (e.g., 'Use for reproducible bugs, failing tests, or unexpected behavior') rather than relying on the separate trigger field, which would raise completeness and trigger-term coverage.

Include natural synonyms users say when reaching for this skill — bug, error, failing test, crash, regression — instead of jargon like 'bounded feedback loop' and 'original-scenario verification'.

Name 2-3 concrete actions the skill performs (reproduce the symptom, isolate a minimal failing case, verify the fix) so specificity goes beyond a single qualified action.

Dimension	Reasoning	Score
Specificity	Names the domain ('Debug a reproducible symptom') plus one concrete action, but 'bounded feedback loop' and 'original-scenario verification' are methodology qualifiers rather than the several distinct actions anchor 4 requires.	3 / 5
Completeness	The 'what' is stated clearly, but the description contains no 'Use when...' clause or equivalent explicit trigger guidance — the rubric caps completeness at 3 for this (the trigger text lives in a separate frontmatter field, not the description).	3 / 5
Trigger Term Quality	'Debug' and 'reproducible symptom' are natural terms a user would say, but common variations and synonyms (bug, error, failing test, crash, regression) are all absent, matching the anchor-3 example of relevant keywords without variations.	3 / 5
Distinctiveness Conflict Risk	Debugging is a clear niche with mostly distinct triggers ('reproducible symptom', 'original-scenario verification'), with only minor overlap risk against closely related testing or code-review skills.	4 / 5
	Total	13 / 20 Passed

Validation

93%

Warnings & errors only

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

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: nyldn/claude-octopus
Path: .claude/skills/skill-debug/SKILL.md
Commit: 0093dab

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

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.