DataArt/artisyn-implementation-harness / skills/implementation-harness
implementation-harness
FlaggedInstall the portable AI-dev harness, rebuild its operational graph standalone, harvest project skill improvements into its canonical templates, package direct canonical edits into a portable bundle, or apply canonical updates to an existing project. Use /implementation-harness launch, /implementation-harness reindex-graph, or /implementation-harness patch harvest|prepare|apply.
canonical/skills/design/references/03-frame.md
View on GitHub# Frame stage
Apply `{{provider:skills_root}}/design/patterns/metrics.md`; this stage's metrics
unit is `--stage frame`. First action: `stage start --stage frame`.
Resume detection and stage tracking live in `{{provider:skills_root}}/design/SKILL.md`.
`$DESIGN_DIR` comes from the router's Resume Detection. Frame frames the
problem, frames the requirements (evaluation criteria for Evaluate), and frames
the boundaries before any research begins โ using the context Context Gathering
already resolved and recorded in `design/Report.md`'s `## Context` section.
Output: `design/00 Frame.md`.
> โ ๏ธ Follow every step in order. Do not skip, summarize, or improvise. If you are unsure, ask before deviating.
> ๐ **Session continuity:** apply `{{provider:skills_root}}/design/patterns/session-continuity.md`.
## Step 1 โ Write 00 Frame.md
Using `{{provider:skills_root}}/design/templates/frame.md`, write `design/00 Frame.md`. Pull context from `design/Report.md`'s `## Context` and `## Evidence` sections โ the tracker fetch supplies the Known ECs and background; the resolved design context supplies constraints and precedent.
**Problem** โ one paragraph from the ticket description (or `{{provider:input}}`): what is broken or missing, and why it matters now.
**Success** โ two sub-sections:
- _Known ECs_ โ acceptance criteria and requirements stated in the ticket (verbatim, cleaned up)
- _Proposed ECs_ โ requirements the developer didn't think of. Look for gaps: missing edge cases, non-functional requirements (performance, security, GDPR), operational concerns (observability, rollback), and future-proofing constraints implied by the architecture or engagement context. Propose 2โ4 additions. Label each `[proposed]` so the developer can accept or reject.
These evaluation criteria become the Requirements ร Options matrix rows during Evaluate.
**Boundaries**
- _In scope_ โ what this design covers
- _Out of scope_ โ what is explicitly excluded (derive from ticket; add obvious exclusions if missing)
**Constraints** โ known deadlines, linked tickets, regulatory requirements, product or architectural constraints that directly bound this decision.
**Sizing** _(optional โ include only when the ticket warrants capacity planning)_
- Expected users / MAU
- Request volume
- Data volume
- Relevant NFR targets (from the resolved design context or stakeholder requirements)
**Links**
- Ticket: `{{PROJECT_KEY}}-{number}`
- Related tickets: `<any linked {{PROJECT_KEY}} tickets if any>`
- Reference docs: `<any {{DESIGN_DOCS_ROOT}} docs referenced>`
Apply the docs-writer authoring pattern (`{{provider:skills_root}}/design/patterns/docs-writer-authoring.md`): delegate the write of `design/00 Frame.md` to `@docs-writer` using the selected delegation procedure โ Decisions is drafted next, in Step 2, not here.
Apply the write-confirm pattern (`{{provider:skills_root}}/design/patterns/write-confirm.md`): say "File saved at `design/00 Frame.md`." Continue straight to Step 2 โ do not stop here.
---
## Step 2 โ Identify decisions
Check the Problem and Success criteria against
`{{provider:skills_root}}/design/patterns/multi-decision-split.md`: does this
bundle two or more choices that could each stand alone โ its own option set,
its own tradeoffs, approvable or reiterable independently of the others?
This is an intelligent-approval checkpoint (`patterns/approval-intelligent.md`).
Auto-approval condition: at most one decision found โ nothing to bundle, so
there is nothing to confirm, and nothing to write.
Condition met โ continue straight to Step 3's sign-off. No `## Decisions`
section is added.
Two or more decisions found โ draft the candidate list (each with why it's
independent) and apply the docs-writer authoring pattern
(`{{provider:skills_root}}/design/patterns/docs-writer-authoring.md`):
delegate adding the `## Decisions` section to `design/00 Frame.md` to
`@docs-writer`. Then this is a sign-off gate (`patterns/approval-sign-off.md`),
on its own โ this is a material structural choice, not a content detail,
since it decides how many times Research through Solution loop:
> "This looks like it bundles {K} independent decisions: {list, each with why it's independent}. Each would get its own Research โ Evaluate โ Solution loop โ confirm the breakdown, merge any, or say it's really one decision."
`wait start --reason approval --detail "Awaiting decisions sign-off"`. **Stop
and wait.** On the reply, delegate any correction to `@docs-writer` the same
way, then record the acceptance check (`--name "Frame decisions sign-off"`,
`--source developer`).
---
## Step 3 โ Sign-off
Apply `patterns/phase-sign-off.md`. Show `design/00 Frame.md` in full if short, or
a summary: Problem, known vs. proposed ECs (with the `[proposed]` ones
called out), Decisions (if Step 2 drafted and already confirmed one), Boundaries, Constraints.
> "Frame drafted at `design/00 Frame.md` โ {N} known requirements, {M} proposed.
> Good to go, or what should change?"
`report`, then `wait start --reason approval --detail "Awaiting frame sign-off"`.
**Stop and wait.** On approval, record the acceptance check
(`--name "Frame sign-off"`, `--source developer`); without it `stage complete` is
refused.
Rework โ apply the named path, present again:
- Source data or problem framing wrong โ "Reiterate from Step 1 โ re-draft the Problem and Boundaries sections."
- Decision breakdown wrong โ "Reiterate from Step 2 โ re-check the Problem/Success criteria and redraft Decisions."
- Target subfolder or design context wrong โ "Reiterate from Context Gathering, Step 3 โ re-resolve context and re-prime."
`design/00 Frame.md` is the single source of ticket context for all subsequent phases. It is not modified again unless the developer explicitly requests a correction.
## Step 4 โ Report
Update `design/Report.md`:
- Phase tracker: `๐ฏ Context Gathering โ
ยท ๐ Ceremony โ
ยท ๐ Frame โ
ยท ๐ Research ๐ ยท ๐ Evaluate โณ ยท ๐ฆ Solution โณ ยท ๐ Review โณ ยท ๐ฌ Address Feedback โณ ยท ๐ค Publish โณ ยท ๐ Completion โณ`
- Append to Efforts table: `| {YYYY-MM-DD} | Frame | {N} known requirements, {M} proposed requirements accepted |`
Then metrics `report`.
## Step 5 โ Commit
Propose:
> `docs({TICKET}): frame <short title>`
This is a manual-operation sign-off (`patterns/approval-sign-off.md`) โ the
developer runs the commit, never you. `wait start --reason user_input`.
**Stop and wait.** On the reply, check `git log`/`git status`: the commit exists
and the worktree holds nothing from this phase. If not, say what is missing and
wait again. Then `stage complete --stage frame`.
๐ Re-read `{{provider:skills_root}}/design/SKILL.md` and resume the router from the next stage.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why โ repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" โ or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself โ your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |