prompt-cosplaySAFE
Most AI agents are markdown in a trench coat. The open-source reality check for agent roles and Agent Skills.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
PROMPT / COSPLAY
Most AI agents are markdown in a trench coat.
The open-source reality check for agent roles and Agent Skills. Separate a polished persona from an execution contract—and an execution contract from proof.
npx --yes github:yangyangnovelist-hub/prompt-cosplay audit github:msitarzewski/agency-agents
PromptCosplay scans SKILL.md files and agent-role Markdown, scores the operational contract, looks for role-specific evals, merges real-use evidence from SkillReaper, and emits a verdict:
It never deletes anything.
The first public autopsy
On msitarzewski/agency-agents@f99f6aa, PromptCosplay detected 283 role-shaped files:
283 roles · 0 SHIP · 253 PROVE · 0 MERGE · 30 BURY HYPE 18/100 · EVIDENCE 62/100 · ~1,010,348 catalog tokens PROOF 0/283 roles have observed use or role-specific evals
That result is more interesting than “the prompts are bad.” They are not: the average operational-contract sco
ad39acc7d9e6OBSERVED · 2026-10-11Install
Commands as the repository documents them. They are shown, not run.
npm install
npm install
Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| codex | mentioned | |
| cursor | mentioned |
What 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: release-gate description: Validate a release before publication. Use when a user asks whether a package is ready to ship. license: MIT --- # Release Gate ## Steps 1. Run the documented test command. 2. Check the package manifest and generated files. 3. Verify that the install command succeeds in a temporary directory. 4. Return a pass/fail report with the exact failing command. ```bash npm test npm pack --dry-run ``` If a command fails, stop and report the failure. Do not publish or push. The gate passes only when all checks succeed.
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 | PASS |
| 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
- declared (2 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (2)
"---\nname: boundary-role\ndescription: Validate proof boundaries.\nevals:\n - .\n - ../../outside.json\n - ../evidence/proof.json\n---\n\nYou are a test role.\n",
commander, skilldoctor, @types/node, tsx, typescript, vitest
Gates applied: no_behavioural_pass.
ad39acc7d9e6full audit observations/trust-audit/skill/yangyangnovelist-hub__prompt-cosplay.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-11 | ad39acc7d9e6 | SAFE | B | 89 | first audit |
Questions
What does the prompt-cosplay skill do?
Most AI agents are markdown in a trench coat. The open-source reality check for agent roles and Agent Skills.
Is prompt-cosplay 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 prompt-cosplay access on my machine?
The audit observed that it runs shell commands. Each of those is consistent with what it says it does. Secrets in the source: none found.
Which assistants does prompt-cosplay work with?
Its documentation mentions codex and cursor. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (ad39acc7d9e6), read on 2026-10-11. The repository is watched, and a new audit runs when it changes — this is the first audit.