Docs WriteSAFE
Intelligent Skill routing and workflow orchestration for AI agents — +21.12 pp reward, −29.6% tokens on SkillsBench with DeepSeekV4Flash-VE.
Overview
Intelligent Skill routing and workflow orchestration for AI agents — +21.12 pp reward, −29.6% tokens on SkillsBench with DeepSeekV4Flash-VE.
5ff3ca429e5bOBSERVED · 2026-10-08What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: docs-write description: Write documentation following Metabase's conversational, clear, and user-focused style. Use when creating or editing documentation files (markdown, MDX, etc.). allowed-tools: Read, Write, Grep, Bash, Glob --- # Documentation Writing Skill @./../_shared/metabase-style-guide.md ## When writing documentation ### Start here 1. **Who is this for?** Match complexity to audience. Don't oversimplify hard things or overcomplicate simple ones. 2. **What do they need?** Get them to the answer fast. Nobody wants to be in docs longer than necessary. 3. **What did you struggle with?** Those common questions you had when learning? Answer them (without literally including the question). ### Writing process **Draft:** - Write out the steps/explanation as you'd tell a colleague - Lead with what to do, then explain why - Use headings that state your point: "Set SAML before adding users" not "SAML configuration timing" **Edit:** - Read aloud. Does it sound like you talking? If it's too formal, simplify. - Cut anything that doesn't directly help the reader - Check each paragraph has one clear purpose - Verify examples actually work (don't give examples that error) **Polish:** - Make links descriptive (never "here") - Backticks only for code/variables, **bold** for UI elements - American spelling, serial commas - Keep images minimal and scoped tight **Format:** - Run prettier on the file after making edits: `yarn prettier --write <file-path>` - This ensures consistent formatting across all documentation ### Common patterns **Instructions:** ```markdown Run: \`\`\` command-to-run \`\`\` Then: \`\`\` next-command \`\`\` This ensures you're getting the latest changes. ``` Not: "(remember to run X before Y...)" buried in a paragraph. **Headings:** - "Use environment variables for configuration" ✅ - "Environment variables" ❌ (too vague) - "How to use environment variables for configuration" ❌ (too wordy) **Links:** - "Check out the [SAML documentation](link)" ✅ - "Read the docs [here](link)" ❌ ### Watch out for - Describing tasks as "easy" (you don't know the reader's context) - Using "we" when talking about Metabase features (use "Metabase" or "it") - Formal language: "utilize", "reference", "offerings" - Too peppy: multiple exclamation points - Burying the action in explanation - Code examples that don't work - Numbers that will become outdated ### Quick reference | Write This | Not This | | -------------------------- | ------------------ | | people, companies | users | | summarize | aggregate | | take a look at | reference | | can't, don't | cannot, do not | | **Filter** button | \`Filter\` button | | Check out [the docs](link) | Click [here](link) |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | NA |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (0)
No findings outside the package's declared scope.
Gates applied: no_behavioural_pass.
5ff3ca429e5bfull audit observations/trust-audit/skill/foryourhealth111-pixel__docs-write.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | 5ff3ca429e5b | SAFE | B | 89 | first audit |
Questions
What does the Docs Write skill do?
Intelligent Skill routing and workflow orchestration for AI agents — +21.12 pp reward, −29.6% tokens on SkillsBench with DeepSeekV4Flash-VE.
Is Docs Write safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Docs Write access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
How current is this page?
The grade is for one exact copy of the source (5ff3ca429e5b), read on 2026-10-08. The repository is watched, and a new audit runs when it changes — this is the first audit.