PokeClaw is an open-source Android app that uses AI to operate a phone, such as reading messages and carrying out tasks. Its local mode runs Gemma 4 on the device without an account or API key, while optional cloud models can help with more difficult reasoning.
Borrowing it
Nothing to install: this file belongs to agents-io/PokeClaw. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/agents-io/PokeClaw/main/CLAUDE.mdgit clone --depth 1 https://github.com/agents-io/PokeClawWrote 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/agents-io/pokeclaw/claude-md)<a href="https://agentmods.dev/instructions/agents-io/pokeclaw/claude-md"><img src="https://agentmods.dev/badge/instructions/agents-io/pokeclaw/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/agents-io/pokeclaw/claude-md"><img src="https://agentmods.dev/badge/instructions/agents-io/pokeclaw/claude-md.svg" alt="Reviewed on agentmods" width="80" 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 | $0.01720 | $0.01720 |
| Opus 5.5 | $0.00688 | $0.00688 |
| Sonnet 5.5 | $0.00344 | $0.00344 |
| Haiku 4.5 | $0.00172 | $0.00172 |
Grade A, and why
PokeClaw 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 2mo 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.
How it starts
The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PokeClaw — Project Rules
Read This First
Before changing prompts, skills, task routing, local runtime, QA, or release behavior, read README.md sections Product Direction, Roadmap, and Known platform constraints.
The short version: PokeClaw is a generic Android mobile-agent harness, not a collection of hardcoded demo tasks. Prioritize runtime, hardware, storage, accessibility, foreground-service, install/signing, and QA-harness correctness before narrow prompt or workflow tuning.
Project Files
| File | What | When to update |
|---|---|---|
README.md |
Public product direction, roadmap, changelog | When product direction, roadmap, release notes, or public claims change |
AI_INDEX.md |
Repo map for coding agents | When files or directories move |
CLAUDE.md |
Project rules | When workflow/rules change |
QA_CHECKLIST.md |
E2E test cases + debug changelog | Every code change |
RELEASING.md |
Stable signing and release workflow | When the release process changes |
BACKLOG.md |
Features, bugs, ideas with priority | When new items come in or items get done |
ARCHITECTURE_RECONSTRUCTION.md |
Historical architecture/refactor guardrails | When reconstruction phases change |
CLAUDE.local.md |
Current session state | Every session |
When Nicole mentions a feature idea, bug, or "之後要做" item → write it into BACKLOG.md immediately. Don't rely on session memory.
QA-First Development (MANDATORY)
Every code change MUST include E2E QA. No exceptions. This is the highest-priority practice in this project.
QA must reflect the product direction in README.md: do not turn stochastic model performance into one-off prompt drilling. Fix and gate deterministic harness failures first, then measure exploratory model tasks by repeated-trial success rate.
Per-Change QA (every commit)
- Design QA tests FIRST — before writing code, define what the E2E test looks like
- Add tests to
QA_CHECKLIST.md— permanent, under the relevant section (A-K or new section) - Tests must be E2E via ADB — simulate real user behavior:
adb shell input tap/text,adb shell am broadcast, uiautomator dump, logcat verification. No unit tests, no mocks. Control the phone like a user would. - Cover edge cases — happy path + error path + boundary conditions. For every feature, ask: "what if permission is missing?", "what if network drops?", "what if user taps twice?", "what if another task is running?"
- Run the new tests — execute them yourself, verify PASS, record results in the QA Debug Changelog
- Run affected existing tests — any section that could be impacted by the change, re-run those tests
- After a big feature — run the FULL checklist top to bottom (all sections A-K+)
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.
- 2mo ago First seen · 123 lines · 1,720 tokens per session scan A 115198be8413
PokeClaw CLAUDE.md is an instructions file published in the GitHub repository agents-io/PokeClaw (1,140 stars, last pushed 5mo ago), licensed Apache-2.0. It adds 1,720 tokens to every session, about $0.0069 per session on Opus 5.5. 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.
PokeClaw CLAUDE.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| PokeClaw CLAUDE.mdagents-io/PokeClaw | Instructions file | 1.1k | 5mo ago | Apache-2.0 | A | 1,720 | Claude Code |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |