Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add melodic-software/claude-code-plugins --skill audit-comment-residuegit clone --depth 1 https://github.com/melodic-software/claude-code-pluginsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/melodic-software/claude-code-plugins/audit-comment-residue)<a href="https://agentmods.dev/skills/melodic-software/claude-code-plugins/audit-comment-residue"><img src="https://agentmods.dev/badge/skills/melodic-software/claude-code-plugins/audit-comment-residue/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/melodic-software/claude-code-plugins/audit-comment-residue"><img src="https://agentmods.dev/badge/skills/melodic-software/claude-code-plugins/audit-comment-residue.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00096 | $0.02503 |
| Opus 5.5 | $0.00038 | $0.01001 |
| Sonnet 5.5 | $0.00019 | $0.00501 |
| Haiku 4.5 | $0.00010 | $0.00250 |
Grade A, and why
audit-comment-residue scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Pre-computed context
Uncommitted code files (empty = none matched or the probe returned nothing): !${CLAUDE_SKILL_DIR}/scripts/changed-code-files.sh 10 2>/dev/null || echo "(git status unavailable)"
Residue findings (sample): !${CLAUDE_SKILL_DIR}/scripts/detect.sh 2>/dev/null | grep -E '^(Summary total:|Finding shape:)' | head -20 || echo "none"
Purpose
Residue is comment text that only makes sense outside the code's present state: what the code used to be, the plan or session that produced it, asides to the requester, back-references to a ticket, PR or branch, and notes on where a block came from or when it was added. Version control owns history; a comment describes the present. It also flags a workaround comment with no link and no removal condition: nothing in it tells a reader the workaround is temporary, so the next author copies it. This skill surfaces candidates with a treatment and changes nothing. Whether a comment captures what the code cannot (a non-obvious why, a constraint, a contract) is the author's call; the shapes below are the comments that fail it.
Residue shapes and treatments
Detection reads only the comment portion of a line, so residue-shaped words in identifiers or string literals are never flagged. Exact cues and boundaries per shape, for explaining a finding or a missing one: reference/shapes.md.
| Shape | Looks like | Tier | Treatment |
|---|---|---|---|
history-narration |
"used to…", "no longer…", "previously", "renamed from X", "we switched from…", "now returns…" | 1 | Delete. Keep only a load-bearing reason, rewritten as present-tense rationale ("must stay ordered because…") |
plan-reference |
"Task 2 replaces the old…", "as planned", "in this PR/commit/refactor" |
1 | Delete; fold any surviving intent into a present-tense why |
conversational-antecedent |
"per your request", "as you asked", "like you said", "per our discussion" |
1 | Delete; the conversation is invisible to every later reader |
history-narration-weak |
"as before", "has always used", "the old <word>", a bare phase number ("(ci-perf Phase 6b)") |
2 | Review: each cue also opens ordinary prose, so delete the narration or keep a present-tense rewrite |
ticket-pr-residue |
"see PR #45", "dotfiles#647", "from the feature branch", "JIRA-123" |
2 | Review: delete a bare back-reference; a TODO(#issue) tracking open work is sanctioned and not flagged |
unjustified-workaround |
A comment naming a workaround (workaround, work around, works around, working around) with neither a link nor a removal condition: "# Workaround: write twice, the cache drops the first write" |
2 | Review: the workaround is open root-cause work. Fix the cause, or give the comment a link (a URL, an issue or PR number such as #412 or owner/repo#77, an RFC number) or a removal condition (until a version or a release or fix, such as until v5, until Node 22, until #123 lands or until upstream fixes; remove when, drop it once, once ... ships; a bare until, as in retry until it succeeds, is not one). Either one on any line of the same comment run justifies the whole run, and that issue reference is not also reported as ticket-pr-residue |
origin-note |
"ported from the dotfiles profile", "Merged 2026-07-24 from dot_bashrc", "Added 2026-08-10 while wiring telemetry" |
1 | Delete; git owns origin. A dated freshness stamp ("verified 2026-09-03 against v2.1.259") stays; marker comments and license or attribution header blocks are exempt |
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- evals/evals.json 10 KB
- evals/fixtures/origin-notes.sh 699 B runs code
- evals/fixtures/residue-snippet.py 490 B runs code
- reference/shapes.md 2.1 KB
- scripts/added-lines.sh 415 B runs code
- scripts/changed-code-files.sh 881 B runs code
- scripts/detect.sh 12 KB runs code
- scripts/detect.test.sh 50 KB runs code
- scripts/lib/comment-shapes.sh 12 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago Changed · +10 lines · +10 tokens per session c249c3b6e446
- 10d ago Changed · -5 lines · -15 tokens per session 10e024ae48ae
- 18d ago Changed · +4 lines d2c3eaa41ed8
- 19d ago Changed · -88 tokens per session fb10f7c2f463
- 28d ago Changed · +3 lines · +30 tokens per session b6affa79c62d
- 1mo ago Changed · +7 lines 9c1172450da1
- 1mo ago Changed · +10 lines 3ac9acda4b3c
- 1mo ago Changed 79459417dc02
- 1mo ago First seen · 104 lines · 159 tokens per session scan A ec537e3a6594
audit-comment-residue is a skill published in the GitHub repository melodic-software/claude-code-plugins (26 stars, last pushed today), licensed MIT. It adds 96 tokens to every session and 2,503 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
refactor-decompose-function
For long functions: break into smaller pieces, extract helper functions, reduce nesting, improve testability and readability.
pr-analyzing
Use this skill when the user asks to analyze a specific GitHub pull request, review a PR's architectural impact, assess what changed in a PR and why it matters, or research the code relationships affected by a PR — example triggers like "analyze PR.
find-shortest-path
For pathfinding and search: shortest path, maze solving, game AI, route planning, graph traversal, BFS/DFS, Dijkstra, A problems.
catch-expected-errors
For iteration with errors: catch exceptions during exploration, skip invalid cases, continue to next attempt.
rust-coding-standards
Rust coding standards, documentation authoring, content workflows, and quality patterns for systemprompt.io development.
visual-design-system
World-class visual design system standards — OKLCH color science, fluid typography, elevation hierarchy, branded asymmetric corners, and dark mode implementation.