$npx -y skills add gtrabanco/agentic-workflow --skill workflow-statusRead-only sensor for orchestrating the workflow programmatically: computes the full state of the project — every feature and fix with its dependency closure (met/unmet), pending fixes, open/merged PRs and their audit state, findings awaiting triage, what is startable right now an
| 1 | # Workflow Status (the orchestrator's sensor) |
| 2 | |
| 3 | One read-only pass over the project that answers, in a single fixed JSON |
| 4 | envelope: **what exists, what is blocked on what, what is startable right now, |
| 5 | and what the recommended next command is.** Built for external orchestrators |
| 6 | (see `docs/workflow/ORCHESTRATION.md`) but equally useful to a human asking |
| 7 | "where do we stand?". |
| 8 | |
| 9 | ## Turn contract — verify before ending the turn |
| 10 | |
| 11 | ``` |
| 12 | ✓ Every claim comes from a RUN command or a READ file (git/forge output, roadmap, |
| 13 | fix index, feature folders) — nothing inferred from memory |
| 14 | ✓ Nothing was edited, committed, pushed, or created — read-only, always |
| 15 | ✓ `next.recommended` is non-bare (carries the unit's slug/NN, never a bare |
| 16 | `/plan-feature`) AND staged by the target unit's resolved status: |
| 17 | `idea`/undesigned → `/design-feature <slug>`; `defined` → `/plan-feature |
| 18 | <slug>`; `planned` → `/execute-phase <NN> P1` |
| 19 | ✓ Every `design_candidates[].next` begins with `/design-feature ` — design |
| 20 | candidates always route to design, regardless of anything else |
| 21 | ✓ When `--last-envelope` is supplied: the no-progress guard ran (crash-recovery |
| 22 | checklist) — a hint that recommended `/plan-feature`/`/design-feature` for a |
| 23 | unit still at its pre-advance status produces a `workflow_observations` note, |
| 24 | never a silently repeated bland recommendation |
| 25 | ✓ `recommendations.product_audit` was computed by the step-16 mechanical |
| 26 | two-condition check (never guessed), and `next.tier` was derived from the |
| 27 | resolved `next.recommended` command via the command→tier map in |
| 28 | `## Machine envelope` (never guessed) |
| 29 | ✓ Per-unit `review`/`closure`/`issues_born` (steps 10–12) were computed per |
| 30 | their fixed rules — `adversarial.ran`/`n` stayed `null` unless real |
| 31 | evidence exists, never guessed — and any fired `next.suggested[]` entries |
| 32 | (step 13) quote their owning skill's condition verbatim, never a second |
| 33 | copy of the trigger logic |
| 34 | ✓ The envelope is emitted on **every** invocation of this skill, including a |
| 35 | same-session natural-language follow-up about state — never replaced by prose |
| 36 | ✓ The emitted envelope was checked against the shape reminders in |
| 37 | `## Machine envelope` (mirroring |
| 38 | `packages/agentic-workflow-schema/envelope.schema.json`) before printing |
| 39 | ✓ The human-readable summary is printed, then the machine envelope (fenced |
| 40 | ```json — see ## Machine envelope) is the ABSOLUTE last output |
| 41 | ``` |
| 42 | |
| 43 | With `--json-only`, skip the human-readable summary: print the envelope alone. |
| 44 | |
| 45 | ## When to use |
| 46 | |
| 47 | - Between orchestration steps: an external driver runs it to decide the next |
| 48 | command and model tier without parsing prose. |
| 49 | - Before picking work manually: "what can I start right now?" |
| 50 | - **Not** for judging quality (that's `review-change`/`audit-pr`) or product |
| 51 | health (that's `product-audit`) — this skill reports state, it never judges. |
| 52 | |
| 53 | ## Step 0 — Discover the project (always first) |
| 54 | |
| 55 | Per the agent guide's **Workflow conventions** + **documentation map**, then |
| 56 | read what THIS skill needs: `docs/features/ROADMAP.md`, the fix index |
| 57 | (`docs/fix/README.md`), every in-flight feature folder's `TASKS.md` + |
| 58 | `progress.md` + `known-issues.md`, and `docs/features/SHIP_DECISIONS.md` if a |
| 59 | ship-roadmap run exists. |
| 60 | |
| 61 | ## Process (fixed sequence — run the commands, don't infer) |
| 62 | |
| 63 | 1. **Git state.** `git branch --show-current`, `git status --porcelain`, |
| 64 | `git fetch` + `git status -sb`. A dirty tree or unpushed branch is reported |
| 65 | as-is (a `workflow`-kind observation in `detail`), never cleaned up. |
| 66 | 2. **Forge state.** List open + recently merged PRs and open issues with the |
| 67 | declared forge CLI (examples use `gh`): |
| 68 | `gh pr list --state open --json number,title,headRefName,url,statusCheckRollup`, |
| 69 | `gh pr list --state merged --limit 20 --json number,headRefName`, |
| 70 | `gh issue list --state open --json number,title,labels`. |
| 71 | 3. **Urgency labels (`detail.urgent`) — labels-only, presence-only, never |
| 72 | decides.** Reuse the open-issue list from step 2 (`gh issue list --json |
| 73 | labels` — the JSON labels array already carried by that call); no separate |
| 74 | call is required. Scan the **labels object only** |