@cobusgreyling/loop-init 1.3.3 → 1.4.0

package/README.md

@@ -4,6 +4,8 @@ Scaffold loop engineering starters into your project by pattern and tool.
 
 **npx @cobusgreyling/loop-init . --pattern daily-triage --tool grok** works immediately.
 
+See [docs/loop-init-validation.md](../../docs/loop-init-validation.md) for a validated pattern × --tool matrix.
+
 ## Install & Run
 
 ```bash
@@ -38,6 +40,10 @@ Fix-capable patterns (`pr-babysitter`, `ci-sweeper`, `dependency-sweeper`, `post
 
 The ledger's `pattern`/`level` let `loop-guard` size the breaker's `--token-budget` from [`loop-cost`](../loop-cost)'s realistic per-run estimate instead of a hand-typed number. The breaker escalates (same error N× in a row, too many consecutive failures, token budget, or iteration cap) instead of looping in vain. Report-only patterns skip it.
 
+Patterns that act on human-authored, often underspecified input (`issue-triage`) also get an **intake** skill:
+
+- `loop-intake` skill — when a work item is too vague to verify "done", it asks one question at a time, pushes for exact values, and writes an open question + `needs-human` escalation instead of guessing. Clarifying up front keeps the loop from burning fix attempts on a goal that was never well defined.
+
 Every scaffold also creates:
 
 - `loop-budget.md` — pattern-specific daily caps and kill switch

package/dist/cli.js

@@ -30,6 +30,12 @@ const PATTERNS_NEEDING_FIX = new Set([
     'dependency-sweeper',
     'post-merge-cleanup',
 ]);
+/**
+ * Patterns that act on human-authored, often underspecified input (issues).
+ * They get the loop-intake skill so the loop clarifies a vague item or escalates
+ * it instead of guessing and burning fix attempts.
+ */
+const PATTERNS_NEEDING_INTAKE = new Set(['issue-triage']);
 const STATE_FILES = {
     'daily-triage': 'STATE.md',
     'pr-babysitter': 'pr-babysitter-state.md',
@@ -190,6 +196,15 @@ async function scaffoldCircuitBreaker(pattern, tool, targetDir, templatesRoot, d
     await writeFile(ledgerPath, seed);
     console.log('  created: loop-ledger.json (circuit breaker)');
 }
+/**
+ * Scaffold the loop-intake skill for patterns that receive ambiguous human
+ * input, so the loop sharpens the goal (or escalates) before acting on it.
+ */
+async function scaffoldIntake(pattern, tool, targetDir, templatesRoot, dryRun) {
+    if (!PATTERNS_NEEDING_INTAKE.has(pattern))
+        return;
+    await copyTemplateSkill(templatesRoot, 'SKILL.md.loop-intake', targetDir, tool, 'loop-intake', dryRun);
+}
 function formatTokenCap(n) {
     if (n >= 1_000_000)
         return `${n / 1_000_000}M`;
@@ -492,6 +507,7 @@ Examples:
     await copyL2Templates(pattern, tool, targetDir, templatesRoot, dryRun);
     await scaffoldCircuitBreaker(pattern, tool, targetDir, templatesRoot, dryRun);
     await scaffoldObservability(pattern, tool, targetDir, templatesRoot, dryRun);
+    await scaffoldIntake(pattern, tool, targetDir, templatesRoot, dryRun);
     await scaffoldConstraints(targetDir, templatesRoot, tool, dryRun);
     if (tool !== 'opencode' && !dryRun && !(await exists(path.join(targetDir, 'AGENTS.md')))) {
         const agentsTemplate = `# AGENTS.md
@@ -529,6 +545,10 @@ npm run lint
         console.log('Circuit breaker wired (loop-guard skill + loop-ledger.json):');
         console.log('  npx @cobusgreyling/loop-context --check --ledger loop-ledger.json');
     }
+    if (PATTERNS_NEEDING_INTAKE.has(pattern)) {
+        console.log('');
+        console.log('Intake wired (loop-intake skill): clarify a vague item or escalate before acting.');
+    }
     console.log('');
     console.log(`First loop (${tool}):`);
     console.log(`  ${firstLoopCommand(pattern, tool)}`);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cobusgreyling/loop-init",
-  "version": "1.3.3",
+  "version": "1.4.0",
   "description": "Scaffold loop engineering patterns and starters into any project. Supports Grok, Claude Code, Codex, Opencode and more. npx @cobusgreyling/loop-init . --pattern daily-triage --tool opencode",
   "type": "module",
   "bin": {

package/templates/SKILL.md.loop-guard

@@ -38,30 +38,27 @@ After every iteration, append what you just tried:
 `outcome` is `success | failure | noop`. Always include `error` on failures —
 that is how the breaker detects a repeated (stagnant) failure.
 
-## Size the token budget from the pattern (once per run)
+## Size the token budget from the pattern
 
-Don't hand-type a token cap — derive it from the pattern's realistic per-run
-cost so the breaker trips on genuine cost blowup, not a made-up number.
-[`loop-cost`](https://github.com/cobusgreyling/loop-engineering/tree/main/tools/loop-cost)
-already computes this from `patterns/registry.yaml`; read the loop's `pattern`
-and `level` straight from the ledger:
+Don't hand-type a token cap — `loop-context` can derive it directly from the
+pattern's realistic per-run cost
+([`loop-cost`](https://github.com/cobusgreyling/loop-engineering/tree/main/tools/loop-cost)
+computes this from `patterns/registry.yaml`) so the breaker trips on genuine
+cost blowup, not a made-up number. Substitute the ledger's own `pattern`/`level`
+(here: `ci-sweeper` / `L2`):
 
 ```bash
-# Substitute the ledger's own pattern/level (here: ci-sweeper / L2).
-BUDGET=$(npx @cobusgreyling/loop-cost --pattern ci-sweeper --level L2 --json \
-  | node -e 'let d="";process.stdin.on("data",c=>d+=c).on("end",()=>process.stdout.write(String(JSON.parse(d).scenarios.realistic.tokensPerRun)))')
+npx @cobusgreyling/loop-context --check --ledger loop-ledger.json \
+  --budget-from-pattern ci-sweeper --budget-level L2
 ```
 
-`scenarios.realistic.tokensPerRun` is a rounded integer, so it feeds
-`--token-budget` directly. The two tools stay independent — this is shell
-wiring, not a code dependency.
-
 ## Before each iteration
 
 1. Append the previous attempt to `loop-ledger.json`.
 2. Run the breaker with the resolved budget:
    ```bash
-   npx @cobusgreyling/loop-context --check --ledger loop-ledger.json --token-budget "$BUDGET"
+   npx @cobusgreyling/loop-context --check --ledger loop-ledger.json \
+     --budget-from-pattern ci-sweeper --budget-level L2
    ```
 3. Act on the exit code:
    - **0** → continue. Optionally trim the next prompt first:
@@ -83,12 +80,13 @@ wiring, not a code dependency.
 - Never widen thresholds just to keep looping — escalation is a feature, not a failure.
 - Never edit the ledger to hide a repeated error; the breaker exists to catch it.
 - Defaults: 3× same error, 5 consecutive failures, 10 iterations. Tune with
-  `--stagnation`, `--no-progress`, `--max-iterations`, `--token-budget`.
+  `--stagnation`, `--no-progress`, `--max-iterations`, or an explicit
+  `--token-budget` (wins over `--budget-from-pattern` if both are given).
 
 ## Interaction with other skills
 
 - `minimal-fix` / `ci-triage` — record each attempt's outcome + error in the ledger.
 - `loop-verifier` — a verifier rejection is a `failure`; log it so repeats trip the breaker.
 - `loop-constraints` — honors "escalate after N attempts"; this skill makes it mechanical.
-- `loop-budget` — the per-run `--token-budget` here comes from loop-cost's
-  realistic estimate; loop-budget.md still governs the *daily* cap across runs.
+- `loop-budget` — the per-run budget here comes from `--budget-from-pattern`;
+  loop-budget.md still governs the *daily* cap across runs.

package/templates/SKILL.md.loop-intake

@@ -0,0 +1,71 @@
+---
+name: loop-intake
+description: >
+  Clarify an ambiguous goal or work item BEFORE the loop acts on it. Runs at the
+  front of a run when the assigned task is underspecified. Ask one question at a
+  time, push for exact values, and escalate instead of guessing.
+user_invocable: true
+---
+
+# Loop Intake - Clarify Before You Act
+
+Autonomous loops waste attempts (and budget) when they guess at a vague goal.
+This skill runs FIRST, before triage or any action skill, whenever the item the
+loop is about to work on is not specific enough to verify "done".
+
+## When to run
+
+Run intake when the goal or work item is missing any of:
+
+- a single, testable definition of done
+- the exact scope (which repo / branch / paths / tickets)
+- concrete values a fix would depend on (error text, thresholds, versions)
+
+If the goal is already specific and verifiable, skip intake and proceed.
+
+## How to clarify (one question at a time)
+
+1. Read the current state file and the item (issue, PR, request) in full. Do NOT
+   re-ask anything already answered there.
+2. Ask the SINGLE most decision-blocking question. Wait for the answer.
+3. Push for an exact value. "It should be fast" becomes "What p95 latency, in
+   ms?". "Add a limit" becomes "How many per minute?". Re-ask once if the answer
+   is still vague.
+4. Stop as soon as the goal is verifiable. Do not interrogate for its own sake.
+
+Ask in behavior terms, not implementation: what the system must do, what the user
+sees, which values matter. Not: which function, table, or endpoint.
+
+## When an answer stays vague
+
+Do not guess. Record it and move on:
+
+- Write the gap as an open question in the state file: `- [ ] OQ: <question>`.
+- Mark the item `needs-human` (escalate per the loop's handoff rule).
+- If NO blocking question can be resolved, escalate the whole item and stop.
+  Acting on a guess burns attempts the circuit breaker should not have to catch.
+
+## Output
+
+Append a short, sharpened goal to the state file (or the item):
+
+```
+Goal (clarified): <one testable sentence>
+Done when: <verifiable condition>
+Open questions: <N> (needs-human) - <list>
+```
+
+Hand back to triage or the action skill only if "Done when" is now verifiable.
+
+## Interaction with other skills
+
+- Runs BEFORE `loop-triage` / `issue-triage` and any action skill.
+- Feeds `loop-verifier`: the "Done when" line becomes the verifier's check.
+- Pairs with `loop-guard` (circuit breaker): intake prevents wasted attempts on a
+  goal that was never well defined; the breaker catches the ones that slip through.
+- Respects `loop-constraints`: never ask for, or act on, denylisted scope.
+
+## Default behavior
+
+Report-only in week one: propose the clarified goal and open questions, but let a
+human confirm before the loop acts on the sharpened goal.