sno-ai/sno-station-skills / skills/R-recursive-self-improvement/rem-reflect/skill
rem-reflect
VerifiedSelf Learning from this machine's Claude Code, Codex, Hermes, OpenClaw and Cursor sessions. Decide skill changes and lessons with accept/reject/tbd; recall lessons at session start.
scripts/trigger.ts
View on GitHubimport { readCalls } from './calls.ts';
import type { CallRecord } from './calls.ts';
import type { Evidence } from './evidence.ts';
import type { Trace } from './harvest.ts';
import { commandId, commandWord, segments } from './verdicts.ts';
// The matchable condition a lesson is bound to: the failed call whose line range holds one of its evidence
// lines. Code computes it at landing; the model never names it. Evidence inside a successful call, or in
// what people said, gives null and the lesson stays on the session-start list.
export interface TriggerMatch { tool: string; command_id: string | null; error_signature: string | null; path_glob: string | null }
// The excerpt is the tail of the output, so its last non-empty line is the one line sure to be whole.
function errorSignature(excerpt: string | null): string | null {
const last = (excerpt ?? '').split('\n').findLast(line => line.trim()) ?? '';
const cleaned = last.replace(/["'`]/g, '').replace(/\b[0-9a-f]{7,}\b/g, '').replace(/(?<![^\s(=:,])\/\S+/g, '')
.replace(/\s+/g, ' ').trim().slice(0, 80);
return cleaned || null;
}
export function triggerMatch(call: CallRecord): TriggerMatch {
const text = call.command_text;
const edit = ['Edit', 'Write', 'edit'].includes(call.tool) ? /\.([A-Za-z0-9]+)$/.exec(call.target ?? '')?.[1] : undefined;
return {
tool: call.tool,
command_id: text === null ? null : commandId(text) ?? segments(text).map(commandWord).find(word => word !== null) ?? null,
error_signature: errorSignature(call.error_excerpt),
path_glob: edit ? `*.${edit}` : null,
};
}
export function triggerMatchFor(trace: Trace, line: number): TriggerMatch | null {
const call = readCalls(trace).find(item => item.failed && item.line_start <= line && line <= item.line_end);
return call ? triggerMatch(call) : null;
}
// The producers' shared step: the first evidence line of the lesson that falls inside a failed call of its trace.
export function lessonTrigger(trace: Trace | undefined, evidence: readonly Evidence[], log: (line: string) => void): TriggerMatch | null {
const id = evidence[0]?.trace_id;
if (!trace) { log(`trigger match: trace ${String(id)} not in the store; lesson lands without one`); return null; }
try {
for (const item of evidence) {
if (item.trace_id !== trace.trace_id) continue;
const match = triggerMatchFor(trace, item.line_start);
if (match) return match;
}
return null;
} catch (error) {
log(`trigger match: ${trace.trace_id}: ${String(error)}; lesson lands without one`);
return null;
}
}
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 | Unverified | 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 |