Management & product · Planning
workflow
/workflow
Use when starting or resuming a feature through the full QRSPI workflow.
For whom
- Author
- ebullient ↗
- Repository
- qrspi-x ↗
- Licence
- Apache-2.0
- Last repository change
- 23.10.2026
Assessment and security
96/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability21 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
- Files in this package
- 2
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
Deprecation
No signals detected. Repository changed on 23.10.2026, without references to retired models or APIs.
Licence
Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.
What it does and when to use it
Use when starting or resuming a feature through the full QRSPI workflow.
Original author description
Use when starting or resuming a feature through the full QRSPI workflow.
How it works
- Core Philosophy
- Quick contract
- Phase map
- Helper status and resume
- Execution modes
- Transition choices
Original SKILL.md headings · 84 lines, 735 words
SKILL.md preview
--- name: workflow description: 'Use when starting or resuming a feature through the full QRSPI workflow.' when_to_use: 'Use for `/qrspi-x:workflow <feature-name>`. For one step, such as a standalone review, use that step''s skill directly.' disable-model-invocation: false compatibility: Node 22+ --- # QRSPI Workflow Orchestrator Guides the complete QRSPI workflow and stops for human approval between steps. Use a step skill directly for a one-off operation; use `qrspi-x:autoloop` only when the human explicitly chooses unattended execution. ## Core Philosophy Code is the source of truth, and humans gate every transition. The workflow may automate navigation, never approval. The canonical runtime contract is [references/runtime.md](references/runtime.md). Read it for role capabilities, isolation, fallback, and evidence requirements; keep orchestration rules here. ## Quick contract 1. Resolve the feature workspace and the next step. 2. Invoke exactly that step's skill with its named inputs. 3. Wait for its artifact, result, or review verdict. 4. Record the transition through the helper when available. 5. Stop and present choices to the human before continuing. Other files under `./qrspi/<feature>/` are not automatic inputs; read one only when the human explicitly names it for the current step. Never read or edit `loop-state.json` or `history.jsonl` by hand. QRSPI artifacts are disposable; product code and on-disk outputs are the source of truth. ## Phase map | Phase | Skill | Purpose | |---|---|---| | Explore (optional) | `qrspi-x:explore` | Survey before intent exists. | | Init | `qrspi-x:init` | Capture feature intent. | | Query ⇄ Research | `qrspi-x:query` ⇄ `qrspi-x:research` | Ask questions, then gather codebase facts. | | Shape (optional) | `qrspi-x:shape` | Compare approaches when direction is unclear. | | Spec → Plan | `qrspi-x:spec` → `qrspi-x:plan` | Define behavior, then create phases. | | Implement → Review | `qrspi-x:implement` → `qrspi-x:review` | Execute approved work, then check it against the spec. | | Autoloop | `qrspi-x:autoloop` | Run Implement → Review unattended for a chosen scope; it never replaces final review. | ## Helper status and resume If `./qrspi/<feature>/` is missing, invoke `qrspi-x:init`. If it exists but `request.md` is missing, route to Init and stop. With the helper, run: ```bash qrspi-x status --feature <feature> --project <path> ``` Surface findings and follow the reported `next` action. During implementation, trust `current.phase` and `current.planProgress`; use the phase skill named above for `start`, `log`, `decision`, or `history`, and use `--help` for exact syntax. If a loop is in progress, hand it to `qrspi-x:autoloop`. Without the helper, use artifacts and markers directly. Read artifacts in workflow order and choose the first incomplete step whose inputs are present. Resume routing is deterministic: - non-empty `## New Questions` in `research.md` → `Query` refinement; - undecided `approach.md` (`## Decision` blank or `None.`) → stop at the Shape gate; - implementation started → first incomplete phase, then first incomplete step; - missing earlier artifact → its producer skill named above; - otherwise → the next step in the phase map. Changes to `request.md`, `queries.md`, or `research.md` invalidate downstream definition and implementation artifacts. Changes to `approach.md` invalidate Spec, Plan, and implementation progress. Run affected steps forward before using stale outputs. ## Execution modes The normal workflow is interactive and gates every transition. `qrspi-x:autoloop` requires the helper, gates at phase or whole-plan boundaries, runs its own helper lifecycle, and still stops before final review. Do not switch modes implicitly. At each transition, announce the current step, invoke its skill, inspect the result, call the helper transition command when available, and wait for the human's choice. Do not duplicate completion or history records owned by the step skill. If the helper is unavailable, keep navigation in the conversation, use phase markers and artifacts for recovery, and apply the direct-review label and scope rules. There is no helper recovery, durable history, or helper-derived diff scope in this mode. ## Transition choices After Init: Query, refine the request, or cancel. After Query: Research, regenerate questions, or cancel. After Research: Query again when `## New Questions` is non-empty; otherwise Shape, Spec, refine Research, or cancel. If Spec is chosen without Shape, helper mode records why with `history add`, not `decision add`. After Shape: record the human's decision under `## Decision`, then choose Spec, return to Query/Research, refine Shape, or cancel. After Spec: Plan, return to Shape or Query/Research, refine Spec, or cancel. After Plan: Implement all, Implement one phase, Autoloop, refine Plan, or cancel. During Implement: continue, checkpoint Review, or stop. After a phase: phase Review, continue to the next phase, or stop. After Review: PASS continues; PASS WITH CONDITIONS or FAIL offers fix/re-implement, plan revision, or spec revision. After final PASS, helper mode may `log park`; offer completion/parking and leave `./qrspi/<feature>/` for the human to clean up.
Files in this package (2)
- SKILL.md 5 KB
- references/runtime.md 2 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · ebullient/qrspi-x | 96 | Apache-2.0 |
| jeremylongshore/tons-of-skills-marketplace | 100 | MIT |
| SimHacker/moollm | 100 | MIT |
| mlophez/polix-ai | 100 | Apache-2.0 |
| stellarie/minimal-agent-kit | 100 | MIT |
| waynesutton/builder-skills | 100 | Apache-2.0 |
| claude-world/director-mode-lite | 100 | MIT |
| DawitLemmaBalcha/agent-workflow | 100 | MIT |
| zhinkgit/embeddedskills | 100 | MIT |
Database snapshot: 13.11.2026 · package version c4f10416883c (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 96spec
Use when defining the behavioral delta for a QRSPI feature. - 96init
Use when starting a QRSPI feature workspace before Query. - 95accelint-qrspi-propose
Use this skill when the user wants to start the formal QRSPI/OpenSpec planning workflow for a ticket, bug, feature request, or proposed product/CLI/app change before implementation. Invoke it when the user wants a spec-driven change package with questions first, factual… - 90team
Use for the Team pipeline only on explicit request. Never infer from ordinary coding work. Runs QRSPI or a leading route. - 91accelint-qrspi-apply
Implement QRSPI-planned OpenSpec changes with intelligent parallelization. Use when the user wants to apply a QRSPI change, implement tasks with parallelization, or says "apply this QRSPI change", "implement with parallelization", "run the parallel slices". This skill is… - 100lean-sdd
Use when starting, resuming or continuing any multi-step build (new project, feature, milestone, "continue next milestone", "resume X").