context-audit

Token-hygiene audit of the always-loaded context stack — CLAUDE.md, AGENTS.md, auto-memory MEMORY.md, and the bootstrap-rendered identity files (SOUL.md, USER.md, ACCESS_POLICY.md, HEARTBEAT.md) or their harness equivalents. Finds redundancy, contradictions, stale content, compression candidates, and skill-extraction candidates; produces a ranked action list sorted by token savings with a risk class per finding. REPORT-ONLY: this skill never edits any audited file. Recommendations for bootstrap-rendered files target the interview answer bank / templates, never the rendered output. Judging routes through `gbrain eval cross-modal` (single cheap model by default; full multi-model panel is explicit opt-in).

Quality

70%

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 ./plugin-variants/gbrain-coding/skills/context-audit/SKILL.md

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 well-structured, actionable skill body with an explicit validation feedback loop and clear navigation, slightly held back by repetition of the report-only message and some motivational prose that could be trimmed.

Suggestions

Consolidate the report-only / rendered-file safety guidance into one section (Contract or Anti-Patterns) and reference it rather than restating it three times to recover tokens.

Tighten or remove the 'What this is' motivational framing, keeping only the operational definition of the auditor role.

Confirm the sibling references (../conventions/*, ../soul-audit/SKILL.md) resolve in the target tree, since progressive_disclosure depends on those paths being real.

Dimension	Reasoning	Score
Conciseness	The body is mostly operational and efficient, but the report-only / rendered-file safety message is restated across the Contract, Anti-Patterns, and Dedup sections, and 'What this is' reads as motivational padding that could be tightened.	3 / 5
Actionability	It provides copy-paste-ready bash (the enumeration loop, a fully-flagged `gbrain eval cross-modal` invocation, and JUDGE resolution from `gbrain config get`) plus a concrete output template — mostly executable, with only minor dependence on the external gbrain toolchain.	4 / 5
Workflow Clarity	A clear five-step procedure (Enumerate → Analyze → Classify → Judge → Deliver) with an explicit validation gate and feedback loop (exit 1 → 'fix the flagged weaknesses … and re-judge'); the report-only/non-mutating nature means the destructive-operation cap does not apply.	5 / 5
Progressive Disclosure	No bundle files exist, but the body is well organized into labeled sections and uses clearly signaled one-level references (e.g. ../conventions/brain-first.md, ../soul-audit/SKILL.md) without nested chaining; minor gaps only because some referenced sibling paths cannot be verified within the skill bundle.	4 / 5
	Total	16 / 20 Passed

Description

67%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 specific, well-scoped description with strong capability detail and clear distinctiveness, weakened by the absence of an explicit 'when to use' clause and natural trigger phrases (which sit in the separate triggers field instead).

Suggestions

Add an explicit 'Use when…' sentence to the description (e.g., 'Use when the user says the context is too big, asks for a context diet, or wants to reduce system-prompt size') so the 'when' is in the description field itself.

Surface a few natural user phrases directly in the description rather than relying solely on the triggers field, to improve trigger-term coverage for routing.

Dimension	Reasoning	Score
Specificity	The description lists multiple concrete actions — 'Finds redundancy, contradictions, stale content, compression candidates, and skill-extraction candidates' and 'produces a ranked action list sorted by token savings with a risk class per finding' — giving comprehensive, specific coverage rather than vague language.	5 / 5
Completeness	It clearly answers 'what' (token-hygiene audit producing a ranked action list) but provides no explicit 'Use when…' trigger guidance; per the rubric a missing 'when' clause caps completeness at 3.	3 / 5
Trigger Term Quality	It contains relevant keywords like 'token-hygiene audit', 'context stack', and 'token savings', but the natural user-facing phrases ('context is too big', 'context diet') live in the separate triggers field rather than the description, so common variations users would actually say are missing.	3 / 5
Distinctiveness Conflict Risk	It carves a clear niche — 'always-loaded context stack', 'REPORT-ONLY', recommendations targeting 'the interview answer bank / templates, never the rendered output' — making it unlikely to fire for the wrong skill.	5 / 5
	Total	16 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 3 suspicious	Warning

	Total	14 / 16 Passed

Repository: garrytan/gbrain
Path: plugin-variants/gbrain-coding/skills/context-audit/SKILL.md
Commit: 43597b1

Reviewed: 1 day 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.