prekuter/dryforge / codex/plugin/skills/ready
ready
VerifiedUnderstand what you mean before anything is built. Takes anything — a one-line idea, notes, a spec, or a mix — reads the project first, asks only what is yours to decide, and writes your intent down for you to approve. Use when the user invokes the `ready` skill. Requires git.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 27.0 KB |
| agents/openai.yaml | config | 233 B |
| references/3-doc-gate.md | markdown | 6.0 KB |
| references/decompose.md | markdown | 11.2 KB |
| references/dependency-calc.md | markdown | 6.1 KB |
| references/elicitation.md | markdown | 18.2 KB |
| references/example-3doc.md | markdown | 7.8 KB |
| references/first-cycle-review.md | markdown | 3.1 KB |
| references/foundation-format.md | markdown | 5.3 KB |
| references/gap-analysis.md | markdown | 4.7 KB |
| references/grounds-gate.md | markdown | 3.6 KB |
| references/intent-completeness.md | markdown | 6.3 KB |
| references/intent-review.md | markdown | 4.0 KB |
| references/output-format.md | markdown | 10.0 KB |
| references/project-design-domain.md | markdown | 4.4 KB |
| references/project-design-technical.md | markdown | 4.2 KB |
| references/project-scoping.md | markdown | 3.5 KB |
| references/review-fidelity.md | markdown | 4.6 KB |
references/grounds-gate.md
View on GitHub# grounds-gate.md — the question filter (shared by DECOMPOSE and ELICIT)
A candidate — a conflict flag from DECOMPOSE, or a probe-generated "unsaid" candidate from ELICIT's
re-measurement — becomes a **confident question** only if it clears this gate. The detectors
deliberately raise more than survives (they are liberal on purpose); this gate is what keeps the
dialogue from degrading into generic "have you thought about concurrency?" spray that fatigues the
user. Shared by both stages so the standard for "worth asking" is one rule, not two.
## The three grounds (state all three, per site)
A candidate is a confident question only if you can state **all three**:
1. **The specific site** — the exact requirement / code path / endpoint / entity it lands on (not
"X in general").
2. **Why it is not already covered** — a *positive* argument that the spec, the plan, the code, the
harness, or a framework/convention default does not already handle it. "Might be missing" /
"unclear" is **not enough**: actively rule out that it is already covered.
3. **The concrete consequence** if it is left unaddressed.
If you cannot ground all three, do **not** raise it as a confident question.
## Asymmetric default
**Insufficient grounds → not a confident question.** This is the inverse of a keep-by-default pass:
escalate only what you can ground. The "why not covered" test (ground 2) is where most candidates
die — it removes the bulk of probe false positives (cross-stack testing showed roughly half of what
the detectors raise are false positives) while keeping the grounded structural/contract gaps.
## A load-bearing-dimension candidate may NOT be *silently* dropped (anti-evasion)
The drop default exists to suppress **noise** — not to let the agent route a load-bearing question it
doesn't want to ask around ELICIT's exit bar (`elicitation.md` — "no guess survives"). That exit only governs
candidates that *became questions*; a candidate killed here never reaches it. So a lazy agent could
under-argue ground 2 ("a framework default probably covers this") to drop a load-bearing dimension
before the bar ever sees it. **Close that route:** when the dropped candidate is on a **load-bearing
dimension the decision-surface accounting raised** (`elicitation.md`), the drop is **not silent** — it must be **recorded in the
spec as that dimension's `N/A — covered by [the ground-2 argument]`** (e.g. "Concurrency: N/A — the
framework's transaction default serializes these writes"). This converts a silent suppression into an
**auditable, falsifiable claim** the independent 3-doc-gate can check against the code. Dropping noise
stays silent; dropping a load-bearing dimension leaves evidence.
## Two tiers (so the gate never silently buries a real gap)
- **Confident questions** — grounded on all three. These are the questions you actually ask (ordered
by leverage, each carrying a recommended default — see `elicitation.md` for the recommend-first
rule).
- **Low-confidence candidates** — failed the "why not covered" test (typically a pre-existing trait,
a framework default, or something derivable from a pattern) but might still matter. Hold them
briefly for a quick scan; do not press them as blocking.
The low-confidence tier recovers the rare real-but-hard-to-ground item the gate would otherwise drop,
so the asymmetric default never *silently* buries a real gap.
## Universality guard
Stack-agnostic. The three grounds and the two tiers apply to any candidate in any stack; what counts
as a "framework default" or a "convention" that already covers a site is discovered from the project
at runtime, never assumed here.
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: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — 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" — 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 — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |