Skip to content
Back to skills

Receipts Or Blocked

ASecurity

Reporting contract for every result the agent reports: each claim carries a receipt (URL, id, SHA, command output) or is labelled BLOCKED / UNKNOWN. Use for every status update, thread reply, PR report, outreach log and scheduled-run summary.

  • 31 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 18, 2026
ai-agentsnode

Security analysis

A100/100

Scanned November 18, 2026

npx -y skills add IgorGanapolsky/ThumbGate --skill receipts-or-blocked --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Receipts Or Blocked?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Receipts Or Blocked
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/igorganapolsky-receipts-or-blocked/badge)](https://www.skillsdirectory.com/skills/igorganapolsky-receipts-or-blocked)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: receipts-or-blocked
description: >
  Reporting contract for every result the agent reports: each claim carries a
  receipt (URL, id, SHA, command output) or is labelled BLOCKED / UNKNOWN.
  Use for every status update, thread reply, PR report, outreach log and
  scheduled-run summary.
---

# Receipts or BLOCKED

Evidence over claims. If there's no receipt, it didn't happen yet.

## 1. States are separate

Never collapse these into one word like "done":

`code written` · `tests passed` · `PR opened` · `queued` · `merged` ·
`deployed` · `outcome verified` · `sent` · `posted`

Report the furthest state you have evidence for, and nothing past it.

## 2. What counts as a receipt

| Claim | Receipt |
|-------|---------|
| tests passed | the command and its pass/fail summary line |
| PR opened | PR URL |
| merged | merge commit SHA from `GET /pulls/N` (`merged: true`) |
| deployed | live `/health` showing the version or `buildSha` (CLAUDE.md gate) |
| sent / posted | message id or post URL from the tool's response |
| queued on the Mac | issue URL or sheet row (`mac-dispatch`) |
| revenue | `node scripts/revenue-status.js` output, never a planning doc |

## 3. Report format

```
<outcome in one line>
- <item>: <state> — <receipt>
- <item>: BLOCKED — <quoted reason> — tried: <routes> — unblocks when: <condition>
- <item>: UNKNOWN — <what would tell us>
```

- **BLOCKED** = we tried and something stopped us. Quote the error. Use the
  `blocked-action-fallback` ladder before writing it.
- **UNKNOWN** = we don't have the data. Never guess a number to fill it.
- **SKIPPED** = deliberately not done (for example the do-not-contact list),
  with the reason.
- **ATTEMPTED** = the call was made but no receipt came back. Not done.

## 4. Never

- Never invent ids, URLs, SHAs, counts, revenue or testimonials.
- Never call your own work verified. Producers don't self-certify consequential
  work; a check, a test run, or another reviewer does.
- Never say "done", "live", "shipped" or "merged" without the receipt in the
  same message.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.