next ยท diff

git:20260819.5b258e7 to git:20260905.c68a71a

16 added, 31 removed. Audit A to A.

---
name: next
- description: Report unfinished brief/propose/ship work in this repo and the exact command to resume each, ranked by what to do now. Use when the user asks "what's next", "what's left", "where do I resume", "/next", "la suite", or to re-orient after a context reset. Reads artifact frontmatter as the source of truth; never implements.
+ description: Report unfinished work and the exact next action from existing artifact states. Never resume or implement automatically.
argument-hint: "[--all] [feature]"
- allowed-tools:
- - Bash
- - Read
- - Glob
---
- <objective>
- Answer "what is left to do and how do I resume it" by scanning this repo's brief/propose artifacts, classifying open vs done from their frontmatter, and emitting the single best next command. Read-only. This skill routes; it never builds (that is ship).
- </objective>
+ # Next
- <how_it_works>
- The state lives in the artifacts themselves, not a separate ledger:
- - brief: `docs/brief/<slug>/brief.md` frontmatter `status` (draft -> ready -> shipped) + `tasks.md` checkboxes for progress.
- - propose: `docs/proposals/.../PROPOSAL.md` frontmatter `status` (Draft -> Review -> Accepted -> shipped).
- - OPEN = brief `ready`/`in_progress` or propose `Accepted`. DONE = `shipped`/`superseded`/`Rejected` (hidden). WIP = still authoring (`draft`/`Review`).
- - `resume_cmd` in frontmatter wins; otherwise it is derived from the path (`/ship <folder-or-PROPOSAL.md>`).
+ Run scripts/scan.cjs from the intended repository, not the launch directory
+ by assumption. --json is for tooling; --all includes completed work. The
+ scanner recognizes shipped markers as terminal.
- Because the board is DERIVED on read, it self-prunes the instant `ship` flips a status. No file to keep in sync. See `references/state-contract.md` for the exact frontmatter fields brief/propose/ship maintain.
- </how_it_works>
+ Report the most useful open action. Do not re-read every artifact body after
+ a sufficient scan. Read a specific artifact only for a requested detail.
+ Distinguish implemented work waiting for user acceptance from work to redo.
+ An old todo or reminder is not evidence that a completed fix needs another
+ diagnosis. Never execute the recommended action without a user request.
- <steps>
- 1. Run the scanner from the repo root. Locate `scripts/scan.cjs` under THIS skill's base directory (shown when the skill loads); when installed as a plugin it is `${CLAUDE_PLUGIN_ROOT}/skills/next/scripts/scan.cjs`:
- `node <this-skill-dir>/scripts/scan.cjs [--all] [<extra-repo-roots>]`
- - default: the OPEN board + the recommended `Next:` command.
- - `--all`: also show in-authoring and done items.
- - `--json`: structured output (for tooling).
- - pass extra repo paths to scan sibling repos in one shot.
- 2. Render the result as-is. Keep it short. Do NOT re-read every artifact body; the frontmatter scan is enough.
- 3. If a feature argument is given, open that artifact (brief.md + tasks.md, or PROPOSAL.md) and report: status, done-vs-remaining tasks, the next unchecked task, and the why from the spec.
- 4. Recommend ONE next action: the top-ranked open item's resume command. Hand off; do not run it.
- </steps>
+ ## Execution policy
- <notes>
- - Stale "Accepted"/"ready" items that are actually already built mean the loop was not closed historically (ship now flips status on finish). Offer a one-time reconcile: set `status: shipped` on the already-done artifacts so the board reflects reality.
- - The same scanner backs the SessionStart hook (`--banner`), so the auto clock-in and `/next` never diverge.
- </notes>
+ Work solo. Ask before any subagent or reviewer, even in auto mode. Explain
+ the independent scope and expected benefit first. No hidden advisor, nested
+ delegation, model retuning, repeated successful checks, or progress spam.
+ Use existing context before asking questions. Stop when the requested result
+ is delivered. User stops and scope changes override pending steps.