evolve-status

スター5

フォーク0

更新日2026年7月7日 13:19

Use when inspecting the health of evolve's learning loop — spool depth, deriver queue, injection-cache age, overlay confidence table, and pending proposals ("evolve status", "is evolve working", "memory loop health", "show overlay confidence", "why is my evolve context stale").

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

ソース

uhstray-io

uhstray-io/huhhb

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

evolve-status — loop observability

The learning loop is mostly invisible by design (hooks + a detached flusher), so this skill is how a user sees it working — or sees why it isn't. Run both commands and present the combined picture:

node "${CLAUDE_PLUGIN_ROOT}/scripts/evolve/honcho_client.ts" status
node "${CLAUDE_PLUGIN_ROOT}/scripts/evolve/overlay.ts" report

How to read it

config source: none — the suite is inert. Nothing is captured, nothing injected; huhhb behaves as if evolve were not installed. Point the user at docs/evolve-plan.md for setup if they expected it on.
mode: local — no server: there is no deriver queue; health is the journal count growing and conclusions being distilled from it. A large journal with few conclusions means /evolve-review hasn't run in a while — suggest it (in local mode it is the only thing that forms conclusions).
spool depth > 0 across multiple sessions — the flusher can't reach Honcho. Check flush.log in the state dir that status prints; observations are safe (at-least-once delivery), just undelivered.
cache age — injection context is cache-first by doctrine; staleness is the accepted trade for zero-latency hooks. Old cache + empty spool means no new learnings, which is normal. Old cache + deep spool means see above.
deriver queue pending stuck > 0 — the Honcho API is up but the deriver worker isn't running (self-host: python -m src.deriver with LLM keys). Writes land; conclusions never form.
overlay table — confidence is earned (min(runs/10,1) × success_rate): an overlay at 0.1 worked once, not "works". Flag anything deprecated, any last_error, and anything unused ~60 days (report --json shows last_used) as candidates for the next /evolve-review to propose archiving (pinned overlays exempt).
pending proposals > 0 — a headless review staged work; offer to run /evolve-review now to approve or reject each.

If the user asks what has been learned (not whether the loop is healthy), that's the evolve skill's cost ladder, not this one.

このリポジトリの他の Skills

同じリポジトリ

repo-kickstart

uhstray-io/huhhb

Use when bootstrapping a repository — greenfield or brownfield — into Uhstray's standard development conventions, OpenSpec, and the two-store memory setup. Idempotent and non-destructive: safe to re-run, never overwrites existing docs. Triggers on "repo kickstart", "bootstrap this repo", "set up our conventions", "conform this repo", "initialize plans/openspec/memory", "initialize this repo for openspec and memory", "set up two-store memory here", "index this repo and create its memory bank".

2026-08-025

memory-onboarding

uhstray-io/huhhb

Use when checking or setting up the four-strata memory system on this machine and in the current project — "is my memory set up right", "memory onboarding", "set up memory", "check my memory setup", after a fresh machine/repo, when repo-kickstart or buhhdy's preflight reports memory degradation, or when .claude/memory / MemPalace / evolve looks misconfigured.

2026-08-025

pr-shepherd

uhstray-io/huhhb

Use when any orchestrating agent's development workflow — buhhdy's Workflow 2 is the canonical caller — has open PRs that need driving from creation through merge and cleanup — the terminal step once PRs are open. Triggers on "shepherd this PR", "drive the PR to merge", "babysit the PRs", "monitor the PR to merge", and post-merge close-out (close linked issues, archive the change, remove the worktree, janitor stale branches in the orchestrator's own prefix (buhhdy/* canonical)). Also load to verify PR-lifecycle preconditions (branch protection, CodeRabbit) before shepherding.

2026-08-025

repo-memory

uhstray-io/huhhb

Use when the user says "remember", "don't forget", "save that", "keep that in mind", "recall", "what do we know about", "look that up", or any request to persist or retrieve project knowledge — stores memories in .claude/memory/ using the Claude Code memory format, committed to git so the whole team shares context.

2026-08-025

core-workflows

uhstray-io/huhhb

Use when starting new planning/research on a project or picking up development against an existing plan — buhhdy's two standard, repeatable sequences: Planning & Research (fresh problem to a validated plan — an OpenSpec change where the repo has adopted the conventions — plus issues) and Development (fanout, audit, docs, and pr-shepherd).

2026-07-215

routing-guide

uhstray-io/huhhb

Reference skill for buhhdy's provider routing logic — the decision tree, model tier table, per-provider strengths, skill-to-provider affinity, and cross-review pairings. Load when reasoning about which provider to use and why, or when a task doesn't clearly match the main routing tree.

2026-07-215

evolve-status — loop observability

The learning loop is mostly invisible by design (hooks + a detached flusher), so this skill is how a user sees it working — or sees why it isn't. Run both commands and present the combined picture:

node "${CLAUDE_PLUGIN_ROOT}/scripts/evolve/honcho_client.ts" status node "${CLAUDE_PLUGIN_ROOT}/scripts/evolve/overlay.ts" report

How to read it

config source: none — the suite is inert. Nothing is captured, nothing injected; huhhb behaves as if evolve were not installed. Point the user at docs/evolve-plan.md for setup if they expected it on.

mode: local — no server: there is no deriver queue; health is the journal count growing and conclusions being distilled from it. A large journal with few conclusions means /evolve-review hasn't run in a while — suggest it (in local mode it is the only thing that forms conclusions).

spool depth > 0 across multiple sessions — the flusher can't reach Honcho. Check flush.log in the state dir that status prints; observations are safe (at-least-once delivery), just undelivered.

cache age — injection context is cache-first by doctrine; staleness is the accepted trade for zero-latency hooks. Old cache + empty spool means no new learnings, which is normal. Old cache + deep spool means see above.

deriver queue pending stuck > 0 — the Honcho API is up but the deriver worker isn't running (self-host: python -m src.deriver with LLM keys). Writes land; conclusions never form.

overlay table — confidence is earned (min(runs/10,1) × success_rate): an overlay at 0.1 worked once, not "works". Flag anything deprecated, any last_error, and anything unused ~60 days (report --json shows last_used) as candidates for the next /evolve-review to propose archiving (pinned overlays exempt).

pending proposals > 0 — a headless review staged work; offer to run /evolve-review now to approve or reject each.

If the user asks what has been learned (not whether the loop is healthy), that's the evolve skill's cost ladder, not this one.

name	evolve-status
description	Use when inspecting the health of evolve's learning loop — spool depth, deriver queue, injection-cache age, overlay confidence table, and pending proposals ("evolve status", "is evolve working", "memory loop health", "show overlay confidence", "why is my evolve context stale").