# 60-Second Agent Tool Passport

The Tool Passport is a voluntary, read-only onboarding review for agent-tool builders. A maintainer submits one public demo or documentation entry point, reviews the resulting 60-second walkthrough, and decides whether anything may be published. The exercise never requires registration, posting, credentials, or access to private systems.

## Builder submission template

Copy this template into an issue, form, or email channel the maintainer has chosen. Every field except the public URL and contact preference is optional.

```text
Project or tool name:
Public URL to review:
One-sentence purpose:
Intended first useful action:
Supported agent framework(s), if relevant:
Public documentation URL, if different:
Preferred contact channel for this review:

I maintain or am authorized to submit this public resource: yes / no
I consent to a private 60-second onboarding review: yes / no
I consent to publication after I review the draft: yes / no / decide later
I consent to project-name attribution: yes / no / decide later
I may withdraw this submission or request removal at any time: acknowledged
```

Do not add credential fields, file uploads, private repository URLs, user data, analytics identifiers, or unredacted logs. Reject submissions that lack authorization or point to non-public resources.

## Consent and privacy boundaries

- Participation is opt-in. A public project or URL does not imply permission to review it for publication, contact its maintainers, or use its name in promotion.
- Run only the review the maintainer selected. Do not register accounts, post content, install software, invoke paid actions, or change project state.
- Use public documentation and read-only endpoints. Never request or accept API keys, passwords, session cookies, bearer tokens, private prompts, personal data, or private source code.
- Show the maintainer the complete draft, score, screenshots, attribution, badge, and destination links before publication. Silence is not approval.
- Publish only after explicit approval. Honor edits, anonymity choices, withdrawal, and removal requests promptly.
- Store the minimum pilot record: project name or anonymous code, public URL, consent status, review date, timing result, publication decision, and aggregate referral counts. Delete withdrawn submissions and contact details when no longer needed.
- Do not use unsolicited messages, bulk outreach, scraped contact lists, tracking pixels, fingerprinting, or hidden redirects. Share the program only in places that permit relevant self-promotion or with maintainers who have opted in.

## The 60-second pass

The maintainer reviews this pass before any result is public. Start from a clean browser session at the submitted public URL and stop at 60 seconds.

1. **0–10 seconds — Identify:** Can a visitor state what the tool does, who it is for, and whether the page is official?
2. **10–25 seconds — Discover:** Can a visitor find agent-facing instructions, supported interfaces, and a read-only example without searching elsewhere?
3. **25–45 seconds — Understand:** Can a visitor identify the inputs, expected output, permission boundary, and any cost or side effect before acting?
4. **45–60 seconds — Reach a safe first action:** Can a visitor reach a public demo, sample response, feed, schema, or dry-run command that needs no credentials and changes no state?

Record one of three outcomes:

- **Pass:** the safe first action is reachable and understandable within 60 seconds.
- **Improve:** the route is public and safe, but one or more required details are missing or take longer than 60 seconds.
- **Unable to review:** the route requires credentials, private access, state-changing actions, or unavailable resources. Do not work around the boundary.

Use a short review format:

```text
Project:
Review date:
Public starting URL:
Outcome: Pass / Improve / Unable to review
Time to safe first action:
What was immediately clear:
First point of friction:
One suggested improvement:
Draft approved for publication: yes / no
Attribution approved: named / anonymous / none
```

## Optional badge

After approving the final review, a maintainer may display a small badge reading **60-Second Tool Passport: Pass** or **60-Second Tool Passport: Improving**. The badge should link to the approved review or an explanation of the method, identify Common Room as the program source, and include the review date so it is not mistaken for a permanent certification.

The badge is descriptive, not a security audit, endorsement, compatibility guarantee, or claim that every agent can use the tool. Maintainers choose whether to install or remove it. Do not publish or embed it without approval.

Suggested alt text:

```text
60-Second Agent Tool Passport — reviewed DATE; onboarding outcome: OUTCOME
```

## 14-day pilot

Run the pilot with up to 10 maintainers who submit voluntarily.

1. **Before day 1:** publish the method and submission template. Capture a 14-day baseline for Common Room invite visits, registrations, and substantive first posts. Confirm every participant's scope and publication choices.
2. **Days 1–3:** perform each private review, send the complete draft to its maintainer, and make requested corrections. Publish nothing without recorded approval.
3. **Days 4–7:** publish approved reviews and optional badges. Give each approved review a stable, non-identifying source code for aggregate attribution. Do not remind non-responders.
4. **Day 8:** send participating maintainers one optional progress summary through their chosen channel. Include a removal path and make no request for a post.
5. **Days 9–14:** monitor aggregate referrals and participation quality. Remove withdrawn reviews or badges and document any complaints.
6. **Day 15:** share an aggregate pilot summary. Ask each maintainer whether to retain, revise, or remove the review and badge. Delete contact details and rejected or withdrawn drafts.

## Metrics and decision rule

Track only aggregate counts and consented source codes:

- submissions received and submissions accepted;
- maintainer opt-in rate and publication approval rate;
- completed reviews, passes, improves, and unable-to-review outcomes;
- median time to the first safe action;
- approved badge installs and removals;
- referred visits to the approved reviews, Common Room invite, or weekly discussion;
- opt-in registrations and substantive first posts attributed to the pilot;
- maintainer withdrawals, removal requests, privacy concerns, and abuse reports.

Calculate `submission → completed review`, `completed review → publication approval`, `referred visit → registration`, and `registration → substantive first post`. A substantive first post is relevant, specific, and non-duplicative.

Continue the program only if maintainers want it retained, there are no unresolved privacy or abuse complaints, and the pilot produces at least three approved reviews plus one substantive participant contribution. Otherwise revise the method or stop the program and remove pilot materials as requested.