evolve-loop CLAUDE.md

Project instructions for an automation loop that repeatedly evaluates and merges software changes. They define required documentation, operating rules, and a stop condition after two consecutive cycles with no merged changes.

In plain words
What is it for?
Use them when operating or changing the evolve loop, batch automation, merge trains, flags, release gates, or related documentation. They direct agents to read the repository's canonical rules before changing loop behavior.
Why use it?
They help prevent an automated coding process from continuing through a broken pipeline without investigating why it produced no usable result.

Instructions file

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/mickeyyaya/evolve-loop/claude-md
Clone the repo
git clone --depth 1 https://github.com/mickeyyaya/evolve-loop
Per session 2,196 This file is loaded in full into every session.
When invoked 2,196 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.02196 $0.02196
Opus 5 $0.01098 $0.01098
Sonnet 5 $0.00439 $0.00439
Haiku 4.5 $0.00220 $0.00220

Measured 2d ago against content hash 756a9964d9c1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

evolve-loop CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 79 lines

How it starts

The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Project Instructions (Claude Code)

Read AGENTS.md first — cross-CLI invariants + the 12 Core Agent Rules. This file is the Claude Code overlay (digest). Full runtime detail — env-var table, operator commands, ship classes, publishing pipeline — lives in docs/operations/runtime-reference.md. Read it before touching loop behavior, flags, gates, or releases. Release notes: CHANGELOG.md. Operating policy (canonical, environment-independent): docs/operations/operating-policy.md — pipeline issues are fixed console-first at maximum reasoning; routing is typed plumbing; salvage before requeue; wiring proofs mandatory. The repo, not session memory, is the source of truth for these rules.

Automation Loop Guardrails

  • When running the evolve loop or any batch/merge-train automation, evaluate output after EVERY cycle. If 2 consecutive cycles produce zero ships (0 merged PRs), STOP the loop immediately and root-cause the pipeline before running another wave.
  • Never let a batch run more than 2 unproductive waves "to see if it self-corrects".
  • ADR-0072 reconciliation: a zero-ship streak is SYSTEM-fail evidence — HALT + P0 applies. This does not conflict with "never stop the queue", which governs task-level failures.

Status Reporting

When reporting on a batch or pipeline run, always include: cycles run, ships/lanes completed (e.g. 2/2), open PR numbers, failing CI job names with failure class, and the specific next action. Do not report "running" or "in progress" as a status without these numbers.

Session conventions

  • Confirm direction first: multi-step/multi-cycle work needs a 3-bullet plan + approval. Single-cycle bug fixes, file-path-specified tasks, and approved-plan tasks are exempt.
  • Output discipline: summaries with file:line refs; >300-line findings go to a markdown file, not chat.
  • Long-running jobs: verify health after launch (exit codes, log tail); checkpoint every cycle so --resume works; surface failures immediately.

Read the full file on GitHub · 79 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 79 lines · 2,196 tokens per session scan A 756a9964d9c1

Subscribe to this mod's changes

evolve-loop CLAUDE.md is an instructions file published in the GitHub repository mickeyyaya/evolve-loop (5 stars, last pushed 2d ago), licensed Apache-2.0. It adds 2,196 tokens to every session, about $0.0110 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.