agentleFS
DailybotHQDailybotHQ/deepworkplan-skill/skills/deepworkplan/status/SKILL.md

deepworkplan-status

Skill23 starsChanged 5 months agoMITView on GitHub

Report Lite or Full Deep Work Plan status — format, approval, readiness, progress, checkpoint, blockers and Markdown/state consistency — without modifying anything.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 2 are distinctive, and 9 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 9

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No parameter → interactive scope selection (Step 1).

    line 28
  2. Very specific:

    Use ../shared/plan_paths.py --plans-dir <dwp_dir>/plans list to list plans.

    line 64
  3. Distinctive:

    If a slug is ambiguous, request the full name or ID.

    line 72
  4. Very specific:

    State first: If state.json exists, read it: status, completed_count / task_count, checkpoint, blocked, updated_at/updated_by, and the …

    line 97
  5. Very specific:

    README index: Read the README's goal, Plan Status line, task list ([x]/[ ], any (re-validate…) markers), Standard: line and pre-approval …

    line 101
  6. Very specific:

    Targeted detail only: Read the last completed task's Completion & Log and the current [ ] task file's title/goal.

    line 111
  7. Very specific:

    Enabled addons (any generation, read-only): Run python3 ../shared/config.py enabled --plan <dir> once per repository: it prints the addon …

    line 116
  8. Very specific:

    consistency findings: if any, with the suggested action (resume to reconcile, refine to fix structure).

    line 132
  9. Very specific:

    Not started (all [ ], no commits) · In progress (mixed, recent activity) · Complete (all [x]) · Blocked (state.json.blocked set or the …

    line 149
  10. Very specific:

    Read-only: never executes or modifies plans.

    line 155
  11. Very specific:

    No plans exist → report "No plans found in .dwp/plans/".

    line 170

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

DeepWorkPlan — Status

Report the current state of one or more Deep Work Plans — progress, completed tasks, current task, checkpoint, blockers, consistency — without executing or modifying anything, and without loading the whole plan.

Shared resources (read these)

  • ../shared/context.sh — resolve dwp_dir.
  • ../shared/dwp-paths.md — plans at .dwp/plans/<plan>/.
  • Guide (essential — read for this flow): none. Status is a read-only report; it does not load the methodology guide.

Parameter Support

  • /dwp-status {plan_name} — check a specific plan.
  • /dwp-status latest — check the highest numbered plan, or the most recently modified legacy plan.
  • /dwp-status all — check all plans.
  • No parameter → interactive scope selection (Step 1).

Normalize the PLAN_ prefix; validate that single plans exist under .dwp/plans/. If not found, show available plans and ask the user to choose.

Lite-aware status

For v2 plans report format, materialization, approval and promotion separately from execution progress. Read them from state.json; fall back to the README's Approval row when the field is absent, and to pending when neither is present. A ready pending-review Lite proposal is valid but not executable. Resolve inline task locators against unique README anchors and count only canonical task-index checkboxes, never fenced examples. Report a missing or duplicate anchor, pending approval, or promotion marker as a finding; do not repair it in this read-only flow.

Trust boundary (write scope)

This skill is read-only by contract. allowed-tools lists Bash (which Trust Hub treats as write-capable in general) — here it is used exclusively for read-only inspection. Status reads plan folders, progress checkmarks, and state.json, and reports; it performs no writes of any kind.

It MUST NOT: modify tasks, progress, state.json or source files; "fix" a plan while reporting on it; or write anywhere. If the status reveals an inconsistency, report it and point at refine/resume — do not repair it silently.

Workflow

Step 0 — Check for Parameters

all → scope "all", skip to Step 2. Otherwise resolve the supplied full name, ID, unique slug or latest with ../shared/plan_paths.py --plans-dir <dwp_dir>/plans resolve <selector>, set scope "single", then go to Step 2. No parameter → Step 1.

Step 1 — Ask for Scope

Use ../shared/plan_paths.py --plans-dir <dwp_dir>/plans list to list plans; the numbered folders appear in numeric order. latest is the highest plan ID when numbered plans exist, or the most recently modified legacy plan otherwise. Offer: a single plan (by ID or name), all plans, or latest. Accept an ID, name, all, or latest.

Resolve a single selection with ../shared/plan_paths.py --plans-dir <dwp_dir>/plans resolve <selector>; accept full names, IDs and unique slugs. If a slug is ambiguous, request the full name or ID.

Step 2 — Gather Status Information (compact projection first)

v6 plans first: a plan folder with a manifest.json contract pointer, contract.json or a contracts/ chain is v6 — its records are the journal, not state.json. Report from the shipped read-only commands, never by re-deriving by hand: python3 ../shared/ledger.py --plan <dir> inspect (events, one line each, torn tails reported); python3 ../shared/scheduler.py ready <dir> (next dispatch, read-only); python3 ../shared/resources.py --plan <dir> report / routing / hold (envelope posture, incl. any exhaustion hold); python3 ../shared/outcomes.py --plan <dir> receipt (recomputed outcome receipt — it changes when the records change). Do not run project or views.py render here — both write, and this flow stays read-only; the state.json already on disk is a v6 projection and may be quoted with its generated_at stamp, marked as derived. Status findings (torn tail, no approval event, projection disagreement) are reported with their suggested repair, never repaired. For a v7 plan, also run python3 ../shared/ledger.py --plan <dir> delegate observe (read-only) and report each delegation's task, transport, addon and state — a completed delegate is still asserted until a gate observes it. Then continue with the README index and consistency reading below (items 2–5 apply unchanged; item 1's state.json is a v6 projection).

For each (non-v6) plan:

  1. State first. If state.json exists, read it: status, completed_count / task_count, checkpoint, blocked, updated_at/updated_by, and the per-task status list. This is the compact projection; it is enough for the summary.
  2. README index. Read the README's goal, Plan Status line, task list ([x]/[ ], any (re-validate…) markers), **Standard:** line and pre-approval note. When there is no state.json, the README is the only source and the report says so.
  3. Consistency check (../spec/PLAN_STATE.md §5): compare README checkboxes with state.json task statuses and completed_count; compare [x] tasks that carry a commit with git log --oneline -20; note a checkpoint whose task is [x], a blocked entry on a [x] task, or a task_count that differs from the number of task files. Every disagreement is a finding (which tasks, which direction, markdown is authoritative) — never an edit.
  4. Targeted detail only. Read the last completed task's Completion & Log and the current [ ] task file's title/goal; run git status. Do not read every task file or every gate record.
  5. Compute totals and progress %.

Enabled addons (any generation, read-only). Run python3 ../shared/config.py enabled --plan <dir> once per repository: it prints the addon keys the repository's .dwp/config.json (over ~/.dwp/config.json) enables, one per line, and any fail-closed warning on stderr (../spec/CONFIG.md). Report them as one line — addons enabled: none is an ordinary answer, never a finding; the methodology works alone. Never write the file from this flow.

Step 3 — Generate Status Report

Single plan: header with the resolved plan basename and its actual location; standard (and whether pre-approved for unattended execution); goal; progress (total / completed / pending / %); completed tasks; pending tasks; current status (last completed, next task, checkpoint note, uncommitted work, recent commits); enabled addons (one line); blocked (reason, since, needs) if set; consistency findings if any, with the suggested action (resume to reconcile, refine to fix structure); notes from the README; last task log summary; whether an Executive Report was requested/produced.

All plans: a one-line-per-plan summary (progress, status, blocked?, last activity), then the single-plan detail block for each.

Step 4 — Additional Information (optional)

Offer deeper detail for any plan: all task files and status, git diff for uncommitted changes, full last-task completion log, the gate records of one task, any blockers.

Step 5 — Quick Actions

Suggest: resume (/dwp-resume <plan>), execute (/dwp-execute <plan>), refine (/dwp-refine plan <plan>), or open a specific task file.

Status Classifications

Not started (all [ ], no commits) · In progress (mixed, recent activity) · Complete (all [x]) · Blocked (state.json.blocked set or the current task's log has a blocker note) · Stale (no activity for a long time — optional) · Inconsistent (any Step 2.3 finding — reported alongside the classification).

Important Notes

  • Read-only: never executes or modifies plans; git status is reported, not committed. Cross-reference the README task list with state.json and git commits for accuracy; the markdown is authoritative when they disagree.
  • Cheap by design: the compact projection plus the README index answer the question; task files and logs are read only for the targeted detail above.

Error Handling

  • Plan folder doesn't exist → report; list available plans; ask to select.
  • README missing, or README says Plan Status: materializing → report a partial materialization; read manifest.json (intended title and task_count) and the README task list when present and list which intended files exist and which are missing; suggest /dwp-refine (complete or discard) or /dwp-create. Never report it as executable.
  • state.json unreadable or invalid → report it as a finding; fall back to the README; suggest /dwp-resume (which regenerates it).
  • No plans exist → report "No plans found in .dwp/plans/"; suggest /dwp-create.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.