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/filopedraz/pocket/claude-mdgit clone --depth 1 https://github.com/filopedraz/pocketWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/filopedraz/pocket/claude-md)<a href="https://agentmods.dev/instructions/filopedraz/pocket/claude-md"><img src="https://agentmods.dev/badge/instructions/filopedraz/pocket/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.01880 | $0.01880 |
| Opus 5 | $0.00940 | $0.00940 |
| Sonnet 5 | $0.00376 | $0.00376 |
| Haiku 4.5 | $0.00188 | $0.00188 |
Grade A, and why
pocket 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 5d 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.
This is a copy
100% identical to pocket AGENTS.md — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 187 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Notes for any AI agent (Claude Code, Codex, OpenCode, …) running inside this repo.
What this repo is
A pocket — a portable memory store for CLI agents. The whole point: you
read durable knowledge from memory/general.md at the start of every
session, and you can recall past conversations from chats/ regardless of
which agent originally produced them.
If the human runs out of credits on one provider, they switch CLIs. The folder layout doesn't move. The context is still here.
What to do at session start
-
Read
memory/general.md. That's who the human is and how they want to be talked to. Ground every response in it. If the file is missing, runbin/pocket initfirst. -
Don't auto-ingest
chats/. It can be large. Pull from it on demand viabin/pocketwhen a topic from a past conversation is relevant.
The pocket CLI
The bin/pocket shim runs a small TypeScript CLI in cli/ (Bun + Biome).
It is the only thing you should use to read chats/ or scaffold local
files.
bin/pocket init scaffold memory/general.md + .gitkeep markers
bin/pocket recent [N] last N sessions across all agents
bin/pocket find <pattern> grep across every chats/*.jsonl
bin/pocket show <sessionId> pretty-print one session
bin/pocket tail follow the active session
bin/pocket doctor verify deps + layout
bin/pocket help full list of commands
Files in chats/ follow YYYY-MM-DD_<agent>_<sessionId>.jsonl. Each line
is one hook event the agent emitted (UserPromptSubmit, PostToolUse,
message.updated, etc.) with secrets scrubbed by hooks/log.sh.
The README has the canonical command table — keep it in sync when you add or change a command (see below).
Self-recursive: delegate pocket CLI work to a subagent
The rule, narrowly scoped to work on this repo's bin/pocket CLI:
- If the human asks for a new pocket subcommand and you don't already
know exactly how to write it, spawn a subagent to implement it
instead of researching inside the main chat. The subagent's job is to
return a finished
cli/src/commands/<name>.tsplus the one-line dispatcher patch and the README row. Main context stays small.
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.
- 5d ago First seen · 187 lines · 1,880 tokens per session scan A eafc1a09942e
pocket CLAUDE.md is an instructions file published in the GitHub repository filopedraz/pocket (2 stars, last pushed 3mo ago), licensed MIT. It adds 1,880 tokens to every session, about $0.0094 per session on Opus 5. A static security scan graded it A with 0 findings. It is 100% identical to pocket AGENTS.md, differing in 0 lines, and is treated as a copy.
Other instructions, from other repositories
hindsight CLAUDE.md
Claude Code instructions for vectorize-io/hindsight, covering claude.md, project overview, development commands, local development (api + ui) and start both api server and control plane ui.
souls-directory AGENTS.md
Instructions for thedaviddias/souls-directory, covering agents.md, learned user preferences and learned workspace facts.
aeon CLAUDE.md
Claude Code instructions for aeonfun/aeon, covering aeon, how aeon works, strategy, voice and soul file hierarchy (read in this order).
compendio-mcp AGENTS.md
AGENTS.md instructions for RuloGB/compendio-mcp, covering agents.md, what this is, commands, architecture and mcp tools (progressive disclosure).
agent-knowledge-cards AGENTS.md
AGENTS.md instructions for manojbajaj95/agent-knowledge-cards, covering agents.md, layout, on-disk cards, commands and rules.
knowl CLAUDE.md
Claude Code instructions for dat999zx/knowl, a project described as: Persistent memory for Claude Code, Cursor and Codex. Facts are typed, sourced, and retired when they change. Local SQLite, no API keys.