pi-goal-list-loop-audit 0.29.1 → 0.29.3

package/extensions/goal-loop-forever.ts

@@ -417,7 +417,8 @@ export function auditMeasureCmd(): string {
  * Each iteration: fresh audit pass → append NEW findings → fix the top
  * open ones → check them off with the fix commit. Honesty laws: never
  * fabricate findings, never rewrite the file's history, never check a box
- * without the fix commit existing.
+ * without the fix commit existing (committed with the repo's configured
+ * identity, on the current branch — no invented identities or branches).
  */
 export function auditTarget(): string {
   return `Audit the project for real problems and fix them, iteration by iteration. Every iteration: (1) run a FRESH audit pass over the codebase — spawn Explore subagents for breadth — hunting real issues: bugs, broken flows, regressions, drift between docs and code, dead code, security holes. Not style nits, not speculative refactors. (2) Append every NEW finding as one checkbox line "- [ ] SEVERITY: short description (file:line)" to ${AUDIT_FINDINGS_REL} (create the file on the first finding; append-only — never delete, rewrite, or reorder existing lines; never re-report a finding already listed). (3) Fix the highest-severity OPEN finding(s) — real fixes, committed — then check the box: "- [x] … — fixed in <commit>". (4) Honesty law: never fabricate findings to look busy; never mark a finding fixed without the fix commit existing. When a full audit pass surfaces nothing new AND no open findings remain, say so plainly — the orchestrator counts open findings every iteration and the plateau stop ends the loop when the well is dry.`;

package/extensions/loops/goal.ts

@@ -1602,6 +1602,12 @@ async function showDecisionPrompt(ctx: ExtensionContext): Promise<boolean> {
     if (!pick) return true; // Escape — the widget card remains the fallback
     const idx = options.indexOf(pick);
     const label = g.pauseOptions![idx] ?? pick.replace(/ {2}\(recommended\)$/, "");
+    // v0.29.3: the wipe escape — "… (/glla wipe)" options run the wipe
+    // (which Confirms on its own — destructive actions keep their gate).
+    if (/\(\/glla wipe\)\s*$/.test(label)) {
+      await cmdGllaWipe(ctx);
+      return true;
+    }
     // Executable options — "Label (/goal cancel)" — RUN the command.
     // Placeholder commands (…/<arg>) fall through to the message path.
     const cmdMatch = label.match(/\(\/(goal|list|loop) ([a-z]+)\)\s*$/);
@@ -5031,7 +5037,10 @@ export default function (pi: ExtensionAPI): void {
         if (next.length !== before) changed = true;
       }
       if (changed) pi.setActiveTools(next);
-      if (!toolHealNotified) {
+      if (!toolHealNotified && missing.length > 0) {
+        // v0.29.3: the notify used to fire even when NOTHING was missing —
+        // "glla: 0 agent tool(s) were hidden … re-activated ()" at every pi
+        // start (field-observed in darklord). Warn only on a real heal.
         toolHealNotified = true;
         const list = missing.join(", ");
         ctx.ui.notify(
@@ -5276,7 +5285,12 @@ export default function (pi: ExtensionAPI): void {
       updateGoal({
         status: "paused",
         pauseKind: "decision",
-        pauseOptions: ["Stop the loop, then resume the goal (/loop stop)", "Cancel the goal (/goal cancel) — the loop keeps running"],
+        // v0.29.3: third option — the wipe escape. Old projects stack a
+        // goal AND a loop AND a list from pre-guard versions; arbitrating
+        // between two artifacts the user doesn't even remember is the odd
+        // part — "i feel like wipe does [make sense]". Wipe keeps its own
+        // Confirm (destructive), so picking it is safe to offer.
+        pauseOptions: ["Stop the loop, then resume the goal (/loop stop)", "Cancel the goal (/goal cancel) — the loop keeps running", "Wipe everything — clean slate for stale leftovers (/glla wipe)"],
         pauseRecommended: 1,
         pauseReason: "held on session load — the loop owns the active slot (one active thing at a time)",
         pauseSuggestedAction: "/loop to work the loop, or /loop stop then /goal resume to work the goal",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-goal-list-loop-audit",
-  "version": "0.29.1",
+  "version": "0.29.3",
   "description": "Goal. Loop. Audit. Done. \u2014 a pi-coding-agent extension that supervises long-running work, with isolated auditor on each completion. Beat bamboozling by design: the auditor runs in a fresh session with no extensions, no skills, no editor \u2014 only the read tools needed to verify your goal.",
   "license": "MIT",
   "author": "dracon",

package/prompts/goal-loop-continuation.md

@@ -150,6 +150,7 @@ When the user must CHOOSE between paths, use `pause_goal` with `kind="decision"`
 
 - **Do not modify the objective silently.** The objective is the user's; if it has drifted from what makes sense, use `complete_goal`'s `newObjective` at completion time, or `pause_goal` and propose a `/goal tweak` mid-flight — never just work on something else and claim the original.
 - **Do not pretend completion.** If verification evidence is missing, call `pause_goal` instead of `complete_goal`.
+- **Git discipline: never touch identity or branches.** Commit with the repo's configured identity exactly as-is — no `git config user.*`, no per-commit `git -c user.name=…` overrides, no invented identities like `<task>-agent <…@local>` (field-observed: a phase agent branded itself `phase-e-agent <phase-e@local>` and polluted the history). No creating or switching branches either — commit on the branch you found (usually `main`) and push to its upstream. If git refuses a commit for a missing identity, STOP and ask the user — never invent one.
 - **Do not polish doorknobs.** If you are out of work and the goal is satisfied, call `complete_goal` instead of inventing a side-improvement.
 - **Do not give up early.** If a task is hard, run it down properly. The auditor will catch doorknobs; the agent's job is to do the real work.
 

package/prompts/goal-loop-forever-metricless.md

@@ -45,6 +45,8 @@ ${STRATEGY_NOTE}
 - Never repeat yourself: before acting, check what earlier iterations
   already did (git log / the artifacts themselves) and pick the next
   unaddressed piece of the spec.
+- Git discipline: commit with the repo's configured identity as-is, on the
+  branch you found — never set `git config user.*` or invent an identity.
 - When the spec is genuinely, verifiably exhausted — every section
   addressed, every case handled — say so plainly in your reply instead of
   inventing cosmetic work. The user watches an honest loop gladly; a

package/prompts/goal-loop-forever.md

@@ -44,6 +44,10 @@ ${STRATEGY_NOTE}
 - Do not modify the measure command or anything it reads for configuration;
   gaming the metric is bamboozling and the plateau detector will simply stop
   the loop.
+- Git discipline: commit with the repo's configured identity as-is, on the
+  branch you found — never set `git config user.*`, never use per-commit
+  identity overrides, never invent `<task>-agent <…@local>` identities or
+  scratch branches.
 - The spec is ALIVE: if the target needs sharpening or the metric no longer
   captures "better", call propose_loop_refine with your rationale — the user
   confirms or rejects. Never game or silently abandon the metric; refine it.