Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/martineserios/thebrananpx agentmods add skills/martineserios/thebrana/verify-docsWrote 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/martineserios/thebrana/verify-docs)<a href="https://agentmods.dev/skills/martineserios/thebrana/verify-docs"><img src="https://agentmods.dev/badge/skills/martineserios/thebrana/verify-docs.svg" alt="Measured on agentmods" 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 | $0.01109 |
| Opus 5 | $0.00014 | $0.00554 |
| Sonnet 5 | $0.00006 | $0.00222 |
| Haiku 4.5 | $0.00003 | $0.00111 |
Grade A, and why
verify-docs scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Verify-Docs — Periodic Doc Verification
Run a structural assumption-freshness check + surface a random sample of assumption rows for manual semantic review. No LLM. This is the trigger-evidence collector for t-441 (LLM-assisted drift detection).
When to use
- Quarterly — primary cadence. Run both
--scope docsand--scope claudemd. - Before a major architecture change — sanity-check assumption claims in the docs you're about to revise.
- After a tier backfill — confirm structural staleness clears with the new tiers.
- After a CLAUDE.md edit — run
--scope claudemdto confirm no volatile content was introduced. - On demand — when you suspect a doc says X but code does Y.
Process
1a. Run the structural + sampling check (scope: docs)
Bash: bash system/scripts/verify-docs.sh --sample 5
If you want reproducibility, pass --seed N. For automation/scripting:
Bash: bash system/scripts/verify-docs.sh --json --sample 5 > /tmp/verify-docs-$(date -I).json
1b. Run the CLAUDE.md portfolio scan (scope: claudemd)
Bash: bash system/scripts/verify-docs.sh --scope claudemd
Detects three violation types across all portfolio CLAUDE.md files (excluding worktrees):
- DATED_STATUS — status lines with embedded 202X-MM-DD dates
- PRICING — service cost lines (
ARS N/mes,$N/mes) - TRACKER_TABLE — work-tracker tables (Status + Priority/Effort/Sprint/Assigned)
For JSON output: bash system/scripts/verify-docs.sh --scope claudemd --json
2. Read the sample
Each sample entry shows:
doc— the filetier— effective tier (architecture / methodology / tech) and source (per-row, doc, or default)verified— the row's last_verified dateclaim— the assumption text (truncated to 100 chars; read the full row in the source doc)
For each sample row:
- Read the claim. Open the doc. Read the full row context.
- Check the implementation: does the code/system still behave as the claim says?
- Mark drift:
y(drifted — claim is wrong) orn(still true).
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.
- 3d ago First seen · 117 lines · 28 tokens per session scan A 58d296aaf83d
verify-docs is a skill published in the GitHub repository martineserios/thebrana (3 stars, last pushed yesterday), licensed MIT. It adds 28 tokens to every session and 1,109 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 2026-09-03.
Other skills, from other repositories
joycraft-add-fact
Invoked by session-end or the human after a fact surfaces — route it to the correct context document (production map, dangerous assumptions, decision log, institutional knowledge, troubleshooting).
check-docs-consistency
Cross-reference project docs for drift, stale references, and contradictions. Outputs timestamped report. Discovers Markdown wherever it lives in the repo.
long-doc-governance
A set of rules for managing long documents by checking their length and splitting them when substantial changes would make them too large.
kb-ingest
Ingest documents from the workhub vault inbox into the knowledge base. Classifies notes into projects/knowledge/archive, proposes tasks for actionable items, links, and indexes. Use when new notes land in inbox/ or a path needs filing.
researcher
Use when asked to create research documents, generate PDFs, make one-pagers, or create cheatsheets. Triggers: "research document", "generate PDF", "one-pager", "cheatsheet", "distinctive design". Provides HTML/Markdown/PDF with Terminal, Editorial, Corporate, Industrial, Fresh aesthetic presets.
shipyard-map
Use when performing brownfield analysis on an existing codebase under Codex — onboarding to a new project, generating codebase documentation, or understanding legacy code. Trigger when the user says "map this codebase", "help me understand this code", "document this project", "I just inherited this repo", or runs…