Closes the self-improvement loop: when escape-pit + wedged-jump +
blind fallback all run 3× in a row without freeing the bot, fire a
dedicated proposal at the auto-improver. Pi gets the full context
(journal byKind, last 12 scenario-memory entries, current slim
snapshot) and is asked to either improve escapePit() (dig forward +
down + side, not only up) OR add a brand-new recovery.tunnel-out skill.
- runtime/stuck-incident.js: new checkWedged() path with separate
cooldown (10 min) from the no-progress path. noteResult() ingests
every dispatched action's detail.mode to count wedged completions.
- runtime/bot.js: dispatchAction calls stuckIncident.noteResult(res)
after each result; tick() calls checkWedged() and files the proposal
via writeProposal({editScope:[runtime/actions.js, runtime/skills/,
runtime/reflex.js]}).
This is the architectural piece: a bot wedged in a 1×1 hole now
generates a proposal that Pi can act on (with edit-scope guard rails
+ npm test smoke gate from PR #19), instead of looping wedged-jump
forever.
Verified live: bot now also picks direction from journal —
"explore.far: journal says leanest quadrant=NE → prefer N" — first
time the bot uses persistent memory to choose where to go next.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
244 lines
8.8 KiB
JavaScript
244 lines
8.8 KiB
JavaScript
// Stuck-incident detector. Today's failure-tracker in bot.js fires
|
||
// proposals on repeated *exception-class* failures (TypeError, timeout).
|
||
// This module fires on "no-progress" stagnation — the bot is healthy and
|
||
// the reflex loop hasn't crashed, but a single reason code (e.g.
|
||
// no_food_source, planner_empty) keeps coming back tick after tick.
|
||
//
|
||
// When the same reason persists past STUCK_THRESHOLD_MS we build a
|
||
// proposal body summarising the situation, including:
|
||
// - the no-progress reason
|
||
// - the active milestone + suggested skill
|
||
// - a slim snapshot
|
||
// - the last skill result
|
||
// - per-skill success/failure metrics (if available)
|
||
// - the allowed edit scope ("runtime/skills/<skill>.js" plus tests)
|
||
//
|
||
// The caller (bot.js) is responsible for actually writing the proposal —
|
||
// this module just produces the body string when the threshold trips.
|
||
|
||
const STUCK_THRESHOLD_MS = 5 * 60 * 1000;
|
||
const COOLDOWN_MS = 30 * 60 * 1000;
|
||
|
||
export function createStuckIncidentDetector({ thresholdMs = STUCK_THRESHOLD_MS, cooldownMs = COOLDOWN_MS } = {}) {
|
||
let currentReason = null;
|
||
let firstSeenAt = 0;
|
||
let lastFiredAt = 0;
|
||
|
||
// Additional trigger path: when the bot has produced N "non-productive
|
||
// completions" in a row (same skill returning mode:escape-pit /
|
||
// wedged-jump / blind walk, no inventory or position change), fire a
|
||
// stuck incident immediately — these are exactly the cases where Pi
|
||
// should write a NEW skill, not patch an existing one.
|
||
let consecutiveWedged = 0;
|
||
const WEDGED_FIRE_AT = 3;
|
||
let lastWedgedFireAt = 0;
|
||
const WEDGED_COOLDOWN_MS = 10 * 60_000;
|
||
|
||
function reset() {
|
||
currentReason = null;
|
||
firstSeenAt = 0;
|
||
}
|
||
|
||
// Caller pings this every time an action completes; we look at the
|
||
// mode/detail to decide if it counts as "wedged" (bot didn't really
|
||
// accomplish anything in-world).
|
||
function noteResult(res) {
|
||
const mode = res?.detail?.mode ?? res?.worldDelta?.mode;
|
||
const isWedgedShape = mode === "escape-pit" || mode === "wedged-jump" || mode === "blind" || mode === "wedged-jump";
|
||
if (isWedgedShape && res?.ok) {
|
||
consecutiveWedged++;
|
||
} else if (res?.ok) {
|
||
consecutiveWedged = 0;
|
||
}
|
||
}
|
||
|
||
function wedgedShouldFire(now = Date.now()) {
|
||
if (consecutiveWedged < WEDGED_FIRE_AT) return false;
|
||
if (now - lastWedgedFireAt < WEDGED_COOLDOWN_MS) return false;
|
||
lastWedgedFireAt = now;
|
||
consecutiveWedged = 0;
|
||
return true;
|
||
}
|
||
|
||
// Returns { fire: true, body, kind, editScope } when an incident should
|
||
// be filed this tick; null otherwise.
|
||
function check({ snapshot, lastResult, metrics, journalSummary, scenarioTail, now = Date.now() }) {
|
||
const reason = snapshot?.noProgressReason ?? null;
|
||
if (!reason) {
|
||
reset();
|
||
return null;
|
||
}
|
||
if (reason !== currentReason) {
|
||
currentReason = reason;
|
||
firstSeenAt = now;
|
||
return null;
|
||
}
|
||
if (now - firstSeenAt < thresholdMs) return null;
|
||
if (lastFiredAt > 0 && now - lastFiredAt < cooldownMs) return null;
|
||
lastFiredAt = now;
|
||
|
||
const milestone = snapshot?.curriculum?.milestone;
|
||
const suggested = snapshot?.curriculum?.plan?.skillId;
|
||
const slim = {
|
||
runtimeState: snapshot.runtimeState,
|
||
noProgressReason: reason,
|
||
milestone: milestone?.title ?? null,
|
||
suggestedSkill: suggested ?? null,
|
||
position: snapshot.position,
|
||
health: snapshot.health,
|
||
food: snapshot.food,
|
||
isDay: snapshot.isDay,
|
||
inventoryCounts: Object.keys(snapshot.inventory ?? {}).length,
|
||
closestHostile: snapshot.closestHostile,
|
||
};
|
||
|
||
const editScope = suggested
|
||
? [`runtime/skills/${suggested.replace(/\./g, "-")}.js`, `runtime/skills/`]
|
||
: ["runtime/skills/", "runtime/"];
|
||
|
||
const metricsLine = metrics && Object.keys(metrics).length
|
||
? Object.entries(metrics)
|
||
.map(([id, m]) => `${id}: ok=${m.ok} fail=${m.fail}`)
|
||
.join(", ")
|
||
: "(none)";
|
||
|
||
const journalLine = journalSummary
|
||
? `byKind=${JSON.stringify(journalSummary.byKind ?? {})} buckets=${journalSummary.totalBuckets ?? 0}`
|
||
: "(no journal)";
|
||
|
||
const scenarioLines = Array.isArray(scenarioTail) && scenarioTail.length
|
||
? scenarioTail.map((e) =>
|
||
`- \`${e.skillId}\` ${e.ok ? "OK" : "FAIL"} code=${e.code ?? "?"} ${e.detail ? `(${e.detail.slice?.(0, 120) ?? ""})` : ""} situ=${e.situationHashShort ?? "?"}`,
|
||
).join("\n")
|
||
: "_(no scenario memory recorded yet)_";
|
||
|
||
const body = [
|
||
`# Stuck on \`${reason}\``,
|
||
"",
|
||
`The runtime has reported the same no-progress reason for >${Math.round(thresholdMs / 60000)} min without a productive action.`,
|
||
"",
|
||
"## Current state",
|
||
"",
|
||
"```json",
|
||
JSON.stringify(slim, null, 2),
|
||
"```",
|
||
"",
|
||
"## Last action result",
|
||
"",
|
||
lastResult
|
||
? `\`${lastResult.label}\` → ${lastResult.code ?? (lastResult.ok ? "ok" : "fail")}${lastResult.detail ? ` (${JSON.stringify(lastResult.detail).slice(0, 200)})` : ""}`
|
||
: "_(none recorded)_",
|
||
"",
|
||
"## Skill metrics so far (this process lifetime)",
|
||
"",
|
||
metricsLine,
|
||
"",
|
||
"## World journal (what we have discovered so far)",
|
||
"",
|
||
journalLine,
|
||
"",
|
||
"## Recent scenario memory (last attempts, what worked / failed in similar situations)",
|
||
"",
|
||
scenarioLines,
|
||
"",
|
||
"## Suggested fix",
|
||
"",
|
||
suggested
|
||
? `Improve \`${suggested}\` so the bot can clear the \`${reason}\` blocker, OR teach a NEW skill that handles this kind of situation if no single edit fixes it. Touch only the listed files (the test files under runtime/**/*.test.js are auto-allowed). Use the scenario-memory entries above to avoid re-introducing patterns that already failed.`
|
||
: `The curriculum has no suggested skill for this state. Either teach the curriculum a new milestone OR add a recovery skill that turns this reason code into a productive action. The scenario memory above shows what's been tried.`,
|
||
"",
|
||
"## Edit scope (auto-patch must obey this)",
|
||
"",
|
||
editScope.map((p) => `- ${p}`).join("\n"),
|
||
"",
|
||
"## Forbidden",
|
||
"",
|
||
"- Don't touch `.env`, `state/`, `extensions/`, `tui/` unless the scope above includes them.",
|
||
"- Don't add new npm dependencies.",
|
||
"- Don't change git history (no `--amend`, no `git reset --hard`).",
|
||
"",
|
||
].join("\n");
|
||
|
||
return {
|
||
fire: true,
|
||
kind: `stuck-${reason}`,
|
||
summary: `stuck on ${reason}${milestone ? ` (milestone: ${milestone.title})` : ""}`,
|
||
body,
|
||
editScope,
|
||
};
|
||
}
|
||
|
||
function checkWedged({ snapshot, lastResult, metrics, journalSummary, scenarioTail, now = Date.now() }) {
|
||
if (!wedgedShouldFire(now)) return null;
|
||
const slim = {
|
||
runtimeState: snapshot?.runtimeState,
|
||
noProgressReason: snapshot?.noProgressReason,
|
||
position: snapshot?.position,
|
||
health: snapshot?.health,
|
||
food: snapshot?.food,
|
||
isDay: snapshot?.isDay,
|
||
inventoryCounts: Object.keys(snapshot?.inventory ?? {}).length,
|
||
};
|
||
const metricsLine = metrics && Object.keys(metrics).length
|
||
? Object.entries(metrics).map(([id, m]) => `${id}: ok=${m.ok} fail=${m.fail}`).join(", ")
|
||
: "(none)";
|
||
const journalLine = journalSummary
|
||
? `byKind=${JSON.stringify(journalSummary.byKind ?? {})} buckets=${journalSummary.totalBuckets ?? 0}`
|
||
: "(none)";
|
||
const scenarioLines = Array.isArray(scenarioTail) && scenarioTail.length
|
||
? scenarioTail.map((e) =>
|
||
`- \`${e.skillId}\` ${e.ok ? "OK" : "FAIL"} code=${e.code ?? "?"} ${e.detail ? `(${e.detail.slice?.(0, 120) ?? ""})` : ""} situ=${e.situationHashShort ?? "?"}`,
|
||
).join("\n")
|
||
: "_(no scenario memory)_";
|
||
|
||
const body = [
|
||
`# Wedged — escape-pit cannot extract the bot`,
|
||
"",
|
||
`The bot has produced ${WEDGED_FIRE_AT}+ "wedged-jump / escape-pit / blind" completions in a row.`,
|
||
"In-world it stands still; the existing escape primitives are not enough.",
|
||
"",
|
||
"## Current state",
|
||
"```json",
|
||
JSON.stringify(slim, null, 2),
|
||
"```",
|
||
"",
|
||
"## Last action result",
|
||
lastResult
|
||
? `\`${lastResult.label}\` → ${lastResult.code ?? (lastResult.ok ? "ok" : "fail")} ${lastResult.detail ? `(${JSON.stringify(lastResult.detail).slice(0, 200)})` : ""}`
|
||
: "_(none)_",
|
||
"",
|
||
"## Skill metrics",
|
||
metricsLine,
|
||
"",
|
||
"## World journal byKind",
|
||
journalLine,
|
||
"",
|
||
"## Recent scenario memory (last attempts)",
|
||
scenarioLines,
|
||
"",
|
||
"## Suggested fix",
|
||
"",
|
||
"Either improve `escapePit()` in `runtime/actions.js` (e.g. dig forward + down + side, not only up) OR add a NEW skill `recovery.tunnel-out` that breaks the bot out of a 1×1 hole by digging a 3-block tunnel in the most-free cardinal. Add tests under `runtime/skills/`.",
|
||
"",
|
||
"## Edit scope",
|
||
"- runtime/actions.js",
|
||
"- runtime/skills/",
|
||
"- runtime/reflex.js",
|
||
"",
|
||
"## Forbidden",
|
||
"- Don't touch `.env`, `state/`, `extensions/`, `tui/`, `package.json`.",
|
||
"- Don't add new npm dependencies.",
|
||
].join("\n");
|
||
|
||
return {
|
||
fire: true,
|
||
kind: "wedged-cant-escape",
|
||
summary: "bot wedged in place; escape-pit ran 3× without freeing it",
|
||
body,
|
||
editScope: ["runtime/actions.js", "runtime/skills/", "runtime/reflex.js"],
|
||
};
|
||
}
|
||
|
||
return { check, checkWedged, noteResult, reset };
|
||
}
|