accept
/accept
Judges the open stage by having an independent evaluator subagent drive the real, running build like a user — scenario by scenario, against the rubric — and record an evidence report and the status fields. Use to get a verdict without building anything; `/iyu:drive` dispatches the same evaluator each round.
- Author
- iyulab ↗
- Repository
- claude-plugins ↗
- Licence
- MIT
- Last repository change
- 13.11.2026
Assessment and security
- Files in this package
- 2
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 13.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Judges the open stage by having an independent evaluator subagent drive the real, running build like a user — scenario by scenario, against the rubric — and record an evidence report and the status fields. Use to get a verdict without building anything; `/iyu:drive` dispatches the same evaluator each round.
Arguments
/accept [stage id] [--smoke]
Original author description
Judges the open stage by having an independent evaluator subagent drive the real, running build like a user — scenario by scenario, against the rubric — and record an evidence report and the status fields. Use to get a verdict without building anything; `/iyu:drive` dispatches the same evaluator each round.
How it works
- Process
- Standing rules
SKILL.md preview
---
name: accept
description: Judges the open stage by having an independent evaluator subagent drive the real, running build like a user — scenario by scenario, against the rubric — and record an evidence report and the status fields. Use to get a verdict without building anything; `/iyu:drive` dispatches the same evaluator each round.
argument-hint: "[stage id] [--smoke]"
disable-model-invocation: true
allowed-tools: Read, Glob, Grep, Bash, Agent
---
# Accept — an independent verdict on the open stage
The verdict comes from **a separate subagent** that sees only the contract, the build and a skeptical
brief — an agent judging its own work tends to spot the defect and then approve anyway. This skill
dispatches that subagent and relays its result. Contract, terms and field ownership:
**[stages.md](${CLAUDE_SKILL_DIR}/../_shared/stages.md)**.
## Process
1. Resolve `<root>` ([continuity-docs.md](${CLAUDE_SKILL_DIR}/../_shared/continuity-docs.md) §1) and
read `<root>/stages/stages.json` in the current repository only (a contract elsewhere
belongs to another product). With no contract or no `open` stage, say so, point to
`/iyu:charter`, and end. The stage is the id in `$ARGUMENTS`, else the `open` one.
2. Dispatch **one** `Agent` call (`subagent_type: general-purpose`). Its prompt is the full text of
[references/evaluator.md](${CLAUDE_SKILL_DIR}/references/evaluator.md) followed by absolute paths to
the contract, the stage id and the evidence directory, plus the contract's `run` command and the
mode (`full`, or `smoke` with `--smoke`). Pass only the contract and the build.
3. Read the evidence file it names and relay the verdict in a few lines: each scenario's result, each
rubric score against its threshold, claims that moved, and what could **not** be verified. Report
`unverified` as `unverified`.
## Standing rules
- Your part is dispatching and relaying; the product, the contract's text and the evidence file belong
to others (stages.md §4).
- The evaluator sets `status: passed` only when `owner_accepted` is already true (stages.md §5).
Evaluating before the owner has seen the demo is normal and leaves the stage `open`.
- Re-run after the build changes; an unchanged build gets the same verdict.
Files in this package (2)
- SKILL.md 2 KB
- references/evaluator.md 4 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · iyulab/claude-plugins | 90 | MIT |
| Xyronb1/accept-skill | 100 | MIT |
Database snapshot: 13.11.2026 · package version b4c15f1a2b1e (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
No similar skills in the database.