Dobeu-tech-eco

check-in

Show project progress against the approved end state as a checklist with N of M complete and a percent, from the Sections map, acceptance boxes and open action items. Read-only. Use when asked for a check-in, progress, status, or what is left, or across every session on the machine (`all`).

Dobeu-tech-eco 0 Updated 1w ago
GitHub

Install

npx skillscat add dobeu-tech-eco/dts2-marketplace/check-in

Install via the SkillsCat registry.

SKILL.md

Directory edition: do not run bundled scripts or hooks, a package launcher, or a download piped into a shell. Do not read tokens or credential files from the machine. Carry out the step with Read, Edit, Grep, Glob, and git or gh, and write the files the step names.

Check-in — complete versus outstanding, against the end state

Read-only and runnable any time, mid-run or between runs. It changes no files.

Steps

  1. Find the project. Use the repo root (or the workspace root when python3 "${CLAUDE_PLUGIN_ROOT}/scripts/scope-check.py" . says workspace-member). No .harness/end-state.md → say No end state yet; /dt:setup interviews to one. and stop.
  2. Gather the items the harness already writes; invent no new state format.
    • The ## Sections table and the ## Acceptance criteria boxes in .harness/end-state.md: a section is complete when its status is done (todo, in-progress and blocked are outstanding); an acceptance box is complete when it is [x].
    • Open items from python3 "${CLAUDE_PLUGIN_ROOT}/scripts/action-items.py" <repo> list (secrets, decisions, handoffs, deploy steps): each open one is outstanding, named <id> <group>: <item>.
  3. Summarize and render. Either pass the whole end state to the script, python3 -I "${CLAUDE_PLUGIN_ROOT}/scripts/check-in.py" --from .harness/end-state.md, or, when you also gathered action items, pass everything as one list: python3 -I "${CLAUDE_PLUGIN_ROOT}/scripts/check-in.py" --items '[{"name":"section 1: data model","status":"complete"},{"name":"A1 Secrets: STRIPE_KEY","status":"outstanding"}]'. Add --json for the raw summary (total, complete, outstanding, percent, complete_items, outstanding_items).
  4. Show it exactly as the script prints it: a header N of M complete (P%), then [x] for each complete item and [ ] for each outstanding one. Then add one line naming the next outstanding item and the command that moves it (/dt:run <task>, /dt:action-items, /dt:run resume, or /dt:autopilot resume).
  5. Autopilot. When .harness/autopilot/latest.md exists, print its header line and progress table as written (step, branch, plan vote, build vote, council, PR, CI, merged, status), then one line: Autopilot <stamp>: <done> of <total> steps done, <parked> parked, current: <first step not done or parked>. For each step that went to council, add one line with its tally, the head's decision and the path to .harness/autopilot/council-<stamp>-<step>.md, so the user can read why. A status of stopped adds resume with /dt:autopilot resume. Section statuses in the checklist above already reflect merged steps, because autopilot sets them to done after each merge.
  6. All sessions on this machine. When invoked as /dt:check-in all, or the user asks for every session or all harnesses, run python3 -I "${CLAUDE_PLUGIN_ROOT}/scripts/fleet.py" checkins [--since 24h] [--json] and show its report, leading with "Needs you" (blockers and questions). Then offer, and confirm first, python3 -I "${CLAUDE_PLUGIN_ROOT}/scripts/fleet.py" broadcast --kind checkin --to all to ask every live session for a fresh check-in; re-run checkins a few minutes later. A session answers with fleet.py checkin --session <id> --status "<one line, free text>" [--blocker X ...] [--next N] [--question Q ...]. Sessions answer on their next prompt, tool call or stop; one idle at the prompt answers only when next touched.

Notes

  • An empty list prints 0 of 0 complete (0%); it never errors on a project with no items yet.
  • This skill reports; it never marks anything done. run updates section statuses and acceptance boxes at ship.