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 aiatelie/ai-atelie --skill pr-evidencegit clone --depth 1 https://github.com/aiatelie/ai-atelieWrote 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/aiatelie/ai-atelie/pr-evidence)<a href="https://agentmods.dev/skills/aiatelie/ai-atelie/pr-evidence"><img src="https://agentmods.dev/badge/skills/aiatelie/ai-atelie/pr-evidence/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/aiatelie/ai-atelie/pr-evidence"><img src="https://agentmods.dev/badge/skills/aiatelie/ai-atelie/pr-evidence.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 | $0.00093 | $0.01875 |
| Opus 5 | $0.00046 | $0.00937 |
| Sonnet 5 | $0.00019 | $0.00375 |
| Haiku 4.5 | $0.00009 | $0.00187 |
Grade A, and why
pr-evidence scanned grade A with 1 finding 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 12d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
[ ] 1. Confirm dev server is up: `curl -sf http://localhost:5173/` 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.
pr-evidence
Every AI Atelie PR ships with two evidence tables in the body:
- Baseline — the eight canonical journeys (home loads · home shows demo · create banner · switch model · agent edits canvas · canvas variations · comment translate · cleanup snapshot). Always runs.
- Task evidence — per-PR feature demo. Runs when the contributor passes a
--task <spec>to the runner.
This skill is the contract for invoking the bun run journeys pipeline so the body update lands consistently across PRs.
When to invoke
- Right after
gh pr createlands (immediately after Step 6 ofship-task, formalized as Step 7 there). - When a previous run failed mid-stream and the body is missing the evidence block.
- When the contributor wants to refresh evidence after pushing follow-up commits.
- Trigger phrases: "post evidence", "run journeys", "update PR with videos", "attach the evidence", "refresh PR evidence".
Skip when:
- The PR is docs-only (no
web/orapi/source change). Evidence is wasted minutes and Anthropic-key cost. - The dev server isn't up — fix that first; the runner refuses to start.
- A previous run on the same SHA already populated the block and nothing has changed.
Hard preconditions
bun run devis running on:5173+:5174.- One-time setup done:
bun run setup:attach(Chromium persistent profile at~/.local/state/aiatelie/chromium-profile/). ghCLI authenticated.- Network reach to api.anthropic.com (the agent journeys spend Sonnet calls).
If any precondition fails, STOP and surface the missing piece. Don't paper over.
Workflow
[ ] 1. Confirm dev server is up: `curl -sf http://localhost:5173/`
[ ] 2. Identify the PR number: auto-detected via `gh pr view --json number`,
or `--pr <n>` if running from a different branch.
[ ] 3. Decide whether this PR needs `--task` evidence (see "Task spec
authoring" below). If yes, write a small spec under
`web/tests/e2e/.task-staging/` (or pass an absolute path).
[ ] 4. Run the suite:
- bare `bun run journeys` for the canonical case
- `bun run journeys -- --task <spec> --task-title "X" --task-description "Y"`
when this PR ships a feature worth demoing
[ ] 5. Watch the per-journey progress lines. A failure prints the spec
path and exit code; the runner continues with the rest.
[ ] 6. After the runner reports "PR #N body updated", verify:
`gh pr view <N> --json body --jq .body | grep -c user-attachments`
must return ≥ 16 (8 baseline × 2 artifacts) and ≥ 18 if `--task`
was used.
[ ] 7. Confirm the auto-cleanup ran: `gh api repos/aiatelie/ai-atelie/issues/<N>/comments --jq '.|length'`
must NOT have ballooned (the runner deletes the upload-evidence
claim-comment after the body update).
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.
- 12d ago First seen · 133 lines · 93 tokens per session scan A 10781fdbb426
pr-evidence is a skill published in the GitHub repository aiatelie/ai-atelie (8 stars, last pushed 3mo ago), licensed MIT. It adds 93 tokens to every session and 1,875 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
pptx-html-fidelity-audit
Audit a python-pptx export against its source HTML deck, identify layout/content drift (footer overflow, cropped content, missing italic/em, lost styling, off-rhythm spacing), and re-export with strict footer-rail + cursor-flow layout discipline. Use this skill whenever the user has a .pptx that was generated from an…
orbit-github
Open Orbit briefing skill — selected by the Orbit pipeline when GitHub is the user's only connected connector, or when the user explicitly scopes their daily digest to GitHub. Pulls the past 24 hours of PRs, review requests, issues, CI runs, and merges from the user's authenticated GitHub connection and renders them…
build-test
Run the project's build / typecheck / lint / test commands and emit the build.passing + tests.passing signals devloop convergence reads.
diff-review
Render the patch-edit run's accumulated changes as a reviewable diff, surface it through a GenUI choice surface, and persist the user's accept / reject decision into the artifact manifest.
patch-edit
Apply one rewrite-plan step at a time as small reviewable file edits, never rewriting whole files when a localised change suffices.
critique-theater
Five-dimension design quality review — score the artifact against craft, brand, accessibility, and copy, then fix what falls short before handing it over.