[V3RON]

Closes #620

Status

Implement: done (9/9 green) Audit: complete Review: round 2 passed, 2 findings fixed Hardware: n/a Gate: waiting for the stack (#960 → #969)
Next: stack sync and gate for #157, when the maintainer says so

Done when

• pnpm test:e2e e2e/setup.test.ts passes: the skill lands at the documented place for each scope, and MCP is connected through the tool's command. (e2e/setup.test.ts)
• By hand, on a machine with no Simlock files: simlock setup --tool claude-code, then --tool codex, then starting each tool lists the simlock skill, and asking the agent to use a simulator makes it run simlock whoami and simlock instructions. (Maintainer's manual check; not run.)
• simlock --help lists setup. (src/cli/index.ts USAGE; asserted in src/cli/index.test.ts)
• reference/cli.mdx, reference/errors.mdx and connect-your-agent.mdx describe the command, both tools, both scopes, the paths and the two error codes.
• pnpm check is green. (CI on 79425f2 running; the gate checks it.)

Feature #157 Completion conditions: the CLI, MCP, text-content and website conditions are delivered by #960; the setup conditions (scopes, re-run, refusal, MCP connect, no daemon) by this PR's tests. The by-hand tool check above is still open.

Assumptions

• Assumption: setup needs a filesystem from the CLI environment — CliEnvironment gains filesystem (already a port of buildCliEnvironment) beside the fields the spec lists.
• Assumption: renderSkill lives in src/instructions/skill.ts and is re-exported by index.ts — avoids an import cycle between index.ts and setup.ts.

Review

Spec review: 0 blocking, 0 fixed, 0 notes. Code review: 2 blocking, 2 fixed, 0 notes. Claims review: 0 blocking, 0 fixed, 0 notes.
Rounds: 2. Tokens: 12.2M.
Runs: 9. Cost USD: 5.6963. Agent seconds: 3230.

Assessment: complete

Written by an agent.

[V3RON]

Review ledger

Round 1: 8cfbe3f
Round 2: 79425f2

Runs: 9. Cost USD: 5.6963. Agent seconds: 3230.
Tokens input 380, output 134286, cache read 15477724, cache write 823895.
- build implementer: claude-sonnet-5-5 1251 s $2.1561
- audit implementer: claude-sonnet-5-5 825 s $0.7329
- review 1 reviewer: claude-sonnet-5-5 434 s $0.3583
- review 1 spec: claude-sonnet-5-5 52 s $0.3776
- review 1 code both: claude-opus-5-5 311 s $1.4903
- review 1 claims: claude-sonnet-5-5 45 s $0.1821
- fix 1 implementer: claude-sonnet-5-5 186 s $0.1844
- fix 1 implementer: claude-sonnet-5-5 86 s $0.0923
- review 2 reviewer: claude-sonnet-5-5 40 s $0.1223
Finding outcomes: {"fixed":2}. Classes: {"architecture 10":1,"testing 3":1}.

{"findings":[
{"id":"R1","role":"code","class":"testing 3","outcome":"fixed","path":"src/instructions/setup.ts","round":"1","fix":"277f89373"},
{"id":"R2","role":"code","class":"architecture 10","outcome":"fixed","path":"src/cli/commands/setup.ts","round":"1","fix":"277f89373"}
]}

Written by an agent.