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/legate-dev/telegram-ai-bridge/code-reviewgit clone --depth 1 https://github.com/legate-dev/telegram-ai-bridgeWhat 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 | $0.00767 | $0.00767 |
| Opus 5 | $0.00383 | $0.00383 |
| Sonnet 5 | $0.00153 | $0.00153 |
| Haiku 4.5 | $0.00077 | $0.00077 |
Grade A, and why
telegram-ai-bridge code-review.instructions.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 — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Code Review Instructions
Review posture
You are reviewing code in a Telegram bridge that wraps AI CLI backends. The codebase is ~5.8k LOC source + ~10.7k LOC tests (1.8× test/source ratio), single-user self-hosted deployment, and handles subprocess spawning, SQLite state, and Telegram API interactions. Bugs here cause silent data corruption, broken sessions, or secret leakage to the operator's phone. Treat every change with the seriousness that production deployments on other people's hosts deserve.
Priority checklist
Check these in order. Stop and flag a finding as soon as you see a violation.
1. Contract compliance
- Does the change match
API_CONTRACT.md? If the PR modifies command behavior, binding schema, or backend interfaces, the contract must be updated in the same PR. - Does
sendMessage()signature in backends match whatmessage-handler.jspasses? - Are new columns in
db.jsreflected in bothCREATE TABLEand the migration block?
2. State integrity
- Binding lifecycle: when a chat rebinds to a different CLI (via
/newor inline keyboard), are per-CLI fields (agent, model) explicitly reset tonull? TheCOALESCEpattern insetChatBindingpreserves stale values unless the caller passesnull. - Session cleanup: when sessions are detached or cleaned up, is all associated state cleared?
- Check for values that silently persist across context switches (CLI changes, session rebinds).
3. Input validation
- All Telegram
ctx.match,ctx.message.text, andctx.callbackQuery.dataare untrusted. - Callback query data prefixes (e.g.,
bind:,setmodel:) must be parsed defensively — colons in values, missing parts, unexpected formats. - Directory paths from user input must be validated before use in
execFile.
4. Security
- No secrets (tokens, API keys) in log output or error messages
execFile(notexec) for all subprocess spawning — arguments as array, never string concatenation- SQL uses named parameters (
@param), never string interpolation - New dependencies must be justified and version-pinned
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 · 65 lines · 767 tokens per session scan A 8e9f7f45bf71
telegram-ai-bridge code-review.instructions.md is an instructions file published in the GitHub repository legate-dev/telegram-ai-bridge (1 stars, last pushed 1mo ago), licensed MIT. It adds 767 tokens to every session, about $0.0038 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.
Other instructions, from other repositories
spec-kitty AGENTS.md
Instructions for Priivacy-ai/spec-kitty, covering spec kitty development guidelines, ⚠️ critical: load the project charter first, ⚠️ critical: template source location, ⚠️ critical: use canonical sources, never improvise and ⚠️ critical: git workflow — no direct pushes to origin/main.
parallel-code CLAUDE.md
Instructions for johannesjo/parallel-code, covering parallel code, stack, commands, project structure and conventions.
PokeTokenBar CLAUDE.md
Instructions for chattymin/PokeTokenBar, covering poketokenbar — claude 프로젝트 지침, 참조 문서 (필요할 때 읽는다), 기여 언어 규약 (오픈소스 대비 — english first), 릴리스 (자연어 트리거) and 확장 규약 (새 프로바이더/툴 추가 시).
cc-statistics CLAUDE.md
Instructions for androidZzT/cc-statistics, covering claude.md, project, setup, usage and architecture.
SubFrame AGENTS.md
Instructions for Codename-11/SubFrame, covering subframe - subframe project, core working principle, relationship to native ai tools, session start and concurrent work & worktrees.
stagecraft AGENTS.md
Instructions for telus-labs/stagecraft, covering agents.md, start here (in order), test and lint commands, where things live and load-bearing contracts (do not break without an adr).