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 EthanYoQ/Skill-hub --skill yq-remove-reasoning-leaksgit clone --depth 1 https://github.com/EthanYoQ/Skill-hubWrote 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/ethanyoq/skill-hub/yq-remove-reasoning-leaks)<a href="https://agentmods.dev/skills/ethanyoq/skill-hub/yq-remove-reasoning-leaks"><img src="https://agentmods.dev/badge/skills/ethanyoq/skill-hub/yq-remove-reasoning-leaks/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/ethanyoq/skill-hub/yq-remove-reasoning-leaks"><img src="https://agentmods.dev/badge/skills/ethanyoq/skill-hub/yq-remove-reasoning-leaks.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 | $0.00028 | $0.01460 |
| Opus 5 | $0.00014 | $0.00730 |
| Sonnet 5 | $0.00006 | $0.00292 |
| Haiku 4.5 | $0.00003 | $0.00146 |
Grade A, and why
yq-remove-reasoning-leaks 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 7d 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 — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Trim Reasoning Leakage
Reasoning leakage is durable repository prose written from the authoring session's vantage rather than the project's current, resolvable state. It cites artifacts only that session could see, narrates a change instead of the resulting behavior, argues with a departed reviewer, or walks through reasoning already expressed by code.
The fix is not indiscriminate deletion. Technical prose standard owns the complete-proposition rule: restate every surviving factual clause so it stands at the current revision, then remove the transcript around it. Delete a passage outright only when it carries no durable proposition.
This is guidance, not a script. Pattern searches nominate candidates; semantic review decides them.
The one test
For every suspect passage ask:
Could a reader at the current revision, without the session transcript, review conversation, branch stack, or uncommitted draft, resolve every reference and verify every claim?
If not, restate the surviving facts from the repository's vantage and remove the unresolvable framing. If yes, the passage is not reasoning leakage merely because it mentions history. Current-state surfaces can still contain unnecessary change narration; move durable history to the repository's sanctioned decision, release-note, migration, or postmortem surface.
Taxonomy
- Dead design-session citations —
(decision 7),(audit C2),design §4.7, phase labels, “the design ledger,” or names of drafts not committed anywhere. Cite the real committed owner by name and path, or remove the citation while preserving its factual clause. - Branch, stack, and PR vantage — “a later PR in this stack,” “this PR adds,” “the previous commit.” State the resulting mechanism or extension point. When repository edits are authorized, put pending work in an owned TODO. Create or update an external issue only with explicit authorization; otherwise report the proposed deferral and owner.
- Change narration and indexical version stamps — “used to,” “no longer,” “the old implementation,” “this cut,” “today,” or “now” contrasting with a deleted state. State present behavior. Turn a useful regression history into a present-tense counterfactual such as “without the byte-length guard, multibyte labels double-encode.”
- Review choreography — “rejected in review,” “the reviewer confirmed,” draft ordinals, or round attributions. Keep the decision and rationale in the appropriate decision-record genre; remove who said it and when.
- Reviewer-addressed justification — “the cast is safe—it simply…” or “this is correct because…”. State the invariant that makes the operation safe, or remove the comment when code/types already express it.
- Restatement and derivation transcripts — control-flow narration, obvious proofs, test walkthroughs, and line-by-line previews. Keep only non-obvious contracts, invariants, or assertion rationale.
- Hedges and ownerless planning residue — “probably fine for now,” “should be enough,” vague roadmap claims, and deferrals without an owner. State a measured bound and failure behavior. When repository edits are authorized, add a concrete owned TODO; create or update an external issue only with explicit authorization. Otherwise report the proposed marker, or delete the hedge when it carries no durable proposition.
- Authoring-language slips — working-language fragments, private separators, and draft annotations embedded in prose of another language. Translate the durable fact or remove the residue.
What ships with it
3 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.
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.
- 7d ago First seen · 59 lines · 28 tokens per session scan A 2ae0a2faf2af
yq-remove-reasoning-leaks is a skill published in the GitHub repository EthanYoQ/Skill-hub (9 stars, last pushed 4d ago), licensed MIT. It adds 28 tokens to every session and 1,460 once invoked, about $0.0001 per session on Opus 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 2026-09-04.
Other skills, from other repositories
gws-sheets
Google Sheets: Read and write spreadsheets.
hive.pdf
Read, write, merge, split, rotate, watermark, encrypt, and OCR PDF files using Python (pypdf, pdfplumber, reportlab, pypdfium2) and command-line tools (poppler-utils, qpdf). Use when the user asks to extract text/tables/images from a PDF, create or modify a PDF, combine or split PDFs, OCR a scanned PDF…
gws-docs
Read and write Google Docs.
gws-forms
Read and write Google Forms.
recipe-email-drive-link
Share a Google Drive file and email the link with a message to recipients.
recipe-share-event-materials
Share Google Drive files with all attendees of a Google Calendar event.