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.
npx agentmods add instructions/agenttoolkit/altk-evolve/agents-mdgit clone --depth 1 https://github.com/AgentToolkit/altk-evolveWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01057 | $0.01057 |
| Opus 5 | $0.00528 | $0.00528 |
| Sonnet 5 | $0.00211 | $0.00211 |
| Haiku 4.5 | $0.00106 | $0.00106 |
Grade A, and why
altk-evolve AGENTS.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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
What is Evolve?
Evolve is a Python library and service which enables AI agents to improve through self-reflection.
Key Concepts
- Trajectory: A recorded agent conversation
- Entity: Anything which is appropriately stored in a vector database, such as a guideline, policy, or some other knowledge.
- Namespace: Isolated storage for entities
- Conflict Resolution: LLM-based merging of duplicate/conflicting entities
- Guidelines: Instructions intended to assist an agent in completing some task
Architecture Flow
- Agent completes some task, and the resulting trajectory is automatically saved into a logging framework such as Langfuse or Arize Phoenix.
- The agent can call the sync MCP function, or the user can manually sync, which causes evolve to process the trajectory and save any generated guidelines.
- Generated guidelines are stored as entities with conflict resolution applied.
- Future agents can query the Evolve MCP server to fetch guidelines for similar tasks
Project Directory Tree (Some files omitted for brevity)
.
├── demo (Files used by the Claude Code demo)
│ ├── filesystem
│ └── workdir
├── docs (Data used by README files)
├── explorations (Tangential projects for feeling out future work. Should be avoided unless otherwise prompted.)
│ └── claudecode
├── evolve (Primary Source Root)
│ ├── backend (Entity Database Backend implementations, primarily vector databases)
│ ├── cli (A CLI wrapper over the native Python client)
│ ├── config (All configurations which are derived from environment variables or instantiated as an object)
│ ├── db (A sqlite database for when vector databases are a poor fit for the data)
│ ├── frontend (Interfaces to interact with the backend)
│ │ ├── client (A native Python client which thinly wraps the configured backend)
│ │ └── mcp (An MCP server implementing some high-level methods useful for AI agents)
│ ├── llm (All code that prompts an LLM)
│ ├── schema (All well-defined datatypes used throughout the project)
│ ├── sync (Upstream data sources to be processed and stored in the backend)
│ └── utils (Small reusable code snippets)
├── tests (All tests for evolve)
├── .env.example (Environment variable template file)
└── .env (Environment variables used to configure evolve)
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.
- yesterday First seen · 78 lines · 1,057 tokens per session scan A 50c72ce1e9dd
altk-evolve AGENTS.md is an instructions file published in the GitHub repository AgentToolkit/altk-evolve (105 stars, last pushed 6d ago), licensed Apache-2.0. It adds 1,057 tokens to every session, about $0.0053 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 2026-08-30.
Other instructions, from other repositories
m_flow AGENTS.md
AGENTS.md instructions for FlowElement-xinliuyuansu/m_flow, covering m-flow — developer & agent reference, 1. repository map, extension points, 2. local development and python backend (requires python 3.10 – 3.13).
AgentRecall-X AGENTS.md
AGENTS.md instructions for Goldentrii/AgentRecall-X, covering agentrecall — codex agent instructions, default (5) — always registered, every session, --full adds (1 more, 6 total), trigger phrases → actions and token cost guide.
agidb CLAUDE.md
Instructions for rohansx/agidb, covering agidb — claude code instructions, naming, before changing anything, agent skills and issue tracker.
kungfu copilot-instructions.md
Instructions for kungfu-systems/kungfu: Read and follow AGENTS.md before proposing or running changes. It is the shared agent router for this repository, including the canonical Shifu development and build entrypoint.
amfs CLAUDE.md
Instructions for raia-live/amfs, covering amfs memory — agent instructions, available mcp tools, identity, brain tools (agent-scoped) and shared knowledge tools.
second-brain AGENTS.md
Instructions for LindaHaviv/second-brain, covering instructions for ai coding agents, what to read first, the rules (non-negotiable), common tasks, the sanctioned way and enforcement, not just instructions.