一键在 Manus 中运行任何 Skill

loop-triage

星标2

分支0

更新时间2026年7月3日 07:05

Triage this repo's real work queue at the start of a loop run or session: read rules/milestones.md (session checkpoint + next not-started chunk), the top of rules/completion-log.md, git/CI status, and the MCP brain. Produces a concise prioritized findings report and updates STATE.md. Report-only unless a chunk is authorized.

安装

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

在 Manus 中运行

来源

TerranSyn

TerranSyn/TerranSoulApp

打开 GitHub 仓库查看创作者相关仓库

下载

在 Manus 中运行

Loop Triage — TerranSoulApp

Adapted from starters/minimal-loop-claude/.claude/skills/loop-triage (cobusgreyling/loop-engineering, MIT) to this repo's actual mechanics: the queue is rules/milestones.md, not a CI/issue scan.

Inputs (read in this order)

rules/milestones.md — the session checkpoint section first (merge queue state, machine notes), then the next not-started chunk; resume any in-progress chunk before starting a new one.
Top of rules/completion-log.md — what just shipped (avoid re-doing done work; the completed rows are removed from milestones).
Git/CI status: current branch, git status, and the latest GitHub Actions result on main (a red main outranks every queued chunk).
MCP preflight: brain_health on :7421/:7423/:7422, then brain_search / brain_suggest_context with the task topic BEFORE any broad repo search. Show the receipt. If MCP cannot start, record the blocker — do not silently skip.
STATE.md — the prior run's pointer snapshot.
loop-constraints.md — binding rules (run the loop-constraints skill first if not already loaded).

Output format

Produce a markdown report with these sections, then sync STATE.md:

1. High-Priority (act now)

One line per item: what + why it outranks the queue (red main, floor regression, user mandate) + suggested loop action + effort estimate.

2. Next chunk (the queue's answer)

The chunk id/title from rules/milestones.md, its acceptance criteria, and the worktree/branch it should run in (D:/Git/ts-<name>-wt).

3. Watch (monitor, don't act)

Honest-gap branches, flake watchlist items, dependabot majors.

4. Noise / Ignore

What was looked at and dismissed (keeps the next triage cheap).

5. State updates

Facts the next run must remember → write them into STATE.md (Last run timestamp, item status, human overrides).

Rules

Be brutally concise; signal only — never invent architectural work during triage.
Only mark High-Priority what a maintainer would want to know today.
A done chunk is logged in completion-log.md and its row REMOVED from milestones.md — never leave done rows.
Report-only when no chunk is authorized; the standing continue-order (2026-06-11) authorizes queued chunks but never destructive actions outside the queue.

同仓库更多 Skills

同仓库

ci-guardian

TerranSyn/TerranSoulApp

Never block on CI. After any push to main (or a PR branch), immediately continue other work and hand CI off to a background agent that watches the run and, if it fails, diagnoses → reproduces → fixes → re-gates → re-pushes, looping until the branch is green. Keeps main green without the main loop ever waiting on a ~12-minute CI cycle.

2026-07-092

demo

TerranSyn/TerranSoulApp

Run a TerranSoul feature demo by DRIVING the live app via Playwright over CDP (WebView2 on 127.0.0.1:9222). Use this skill whenever the user types /demo or /demo <scenario> — e.g. /demo jd, /demo resume, /demo e2e, /demo show. Especially trigger for /demo jd, which automates TerranSoul end-to-end: attach 4 resume PDFs, quick-import, and ask in English/Vietnamese/Japanese which candidate fits a role.

2026-07-072

loop-verifier

TerranSyn/TerranSoulApp

Independent checker for loop- or agent-produced changes in this repo. Runs the CI-exact gate (vitest, vue-tsc, eslint, CI-exact clippy, full cargo test --lib) and enforces the never-regress benchmark floors. Default stance REJECT; never implements fixes itself.

2026-07-042

loop-engineering

TerranSyn/TerranSoulApp

Designs and runs agentic feedback loops instead of one-shot prompts. Use it when: (1) starting any coding task an agent will iterate on (bug fix, migration, refactor, integration debug), (2) the user asks to 'design the loop', 'set up validation', 'make the agent run until tests pass', or mentions loop engineering, (3) an agent session is thrashing — repeated edits to the same files with no progress, oversized diffs, or actions taken without observing results, (4) choosing the right feedback signal (tests vs compiler vs review vs runtime logs) for a task, (5) auditing why an autonomous coding loop failed (overfit-to-test, context drift, unsafe autonomy), (6) designing a STANDING loop — a scheduled maintenance loop with a triage skill, state file, worktree isolation, maker/checker split, budget, and human gates (daily triage, CI sweeper, changelog drafter), or running loop-audit/loop-init/loop-cost.

2026-07-032

loop-budget

TerranSyn/TerranSoulApp

Check this repo's resource budgets before and after every loop run: bench-terminal caps, one-bench/one-MCP rules, cargo serial-build limits, and the /loop engine's seeded budgets. Enforces early exit when nothing is actionable and appends a run entry to loop-run-log.md.

2026-07-032

loop-constraints

TerranSyn/TerranSoulApp

Read loop-constraints.md at the start of every loop run and enforce every rule. Runs BEFORE triage or any action skill. Constraints are binding — they encode this repo's real push/merge, denylist-path, bench-purity, and human-gate rules.

2026-07-032

name	loop-triage
description	Triage this repo's real work queue at the start of a loop run or session: read rules/milestones.md (session checkpoint + next not-started chunk), the top of rules/completion-log.md, git/CI status, and the MCP brain. Produces a concise prioritized findings report and updates STATE.md. Report-only unless a chunk is authorized.
user_invocable	true

Loop Triage — TerranSoulApp

Adapted from starters/minimal-loop-claude/.claude/skills/loop-triage (cobusgreyling/loop-engineering, MIT) to this repo's actual mechanics: the queue is rules/milestones.md, not a CI/issue scan.

Inputs (read in this order)

rules/milestones.md — the session checkpoint section first (merge queue state, machine notes), then the next not-started chunk; resume any in-progress chunk before starting a new one.
Top of rules/completion-log.md — what just shipped (avoid re-doing done work; the completed rows are removed from milestones).
Git/CI status: current branch, git status, and the latest GitHub Actions result on main (a red main outranks every queued chunk).
MCP preflight: brain_health on :7421/:7423/:7422, then brain_search / brain_suggest_context with the task topic BEFORE any broad repo search. Show the receipt. If MCP cannot start, record the blocker — do not silently skip.
STATE.md — the prior run's pointer snapshot.
loop-constraints.md — binding rules (run the loop-constraints skill first if not already loaded).

Output format

Produce a markdown report with these sections, then sync STATE.md:

1. High-Priority (act now)

One line per item: what + why it outranks the queue (red main, floor regression, user mandate) + suggested loop action + effort estimate.

2. Next chunk (the queue's answer)

The chunk id/title from rules/milestones.md, its acceptance criteria, and the worktree/branch it should run in (D:/Git/ts-<name>-wt).

3. Watch (monitor, don't act)

Honest-gap branches, flake watchlist items, dependabot majors.

4. Noise / Ignore

What was looked at and dismissed (keeps the next triage cheap).

5. State updates

Facts the next run must remember → write them into STATE.md (Last run timestamp, item status, human overrides).

Rules

Be brutally concise; signal only — never invent architectural work during triage.
Only mark High-Priority what a maintainer would want to know today.
A done chunk is logged in completion-log.md and its row REMOVED from milestones.md — never leave done rows.
Report-only when no chunk is authorized; the standing continue-order (2026-06-11) authorizes queued chunks but never destructive actions outside the queue.