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/pluto2060/ctx/claude-mdgit clone --depth 1 https://github.com/pluto2060/CTXWhat 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 | $0.03854 | $0.03854 |
| Opus 5 | $0.01927 | $0.01927 |
| Sonnet 5 | $0.00771 | $0.00771 |
| Haiku 4.5 | $0.00385 | $0.00385 |
Grade B, and why
CTX CLAUDE.md scanned grade B with 1 finding 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
**single source of truth = `~/.claude/settings.json`의 `hooks` 블록.** 아래 표는 2026-04-19 시점에서 그 파일로부터 검증한 실제 wired hook만 기재. How it starts
The opening of the file, as written. The whole thing — 246 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CTX — Context Bootstrapper for Claude Code (updated 2026-04-19)
새 CTX 정의 — 실제 production 구성 (settings.json 기준)
CTX = Claude Code의 자동 context 주입 시스템.
single source of truth = ~/.claude/settings.json의 hooks 블록. 아래 표는 2026-04-19 시점에서 그 파일로부터 검증한 실제 wired hook만 기재.
현재 wired hooks (production)
| Hook | 파일 | Event | 역할 | 현재 상태 |
|---|---|---|---|---|
| chat-memory | ~/.claude/hooks/chat-memory.py |
UserPromptSubmit | CM: vault.db FTS5 + vec0 hybrid (α=0.5 cosine + 0.5 bm25) — 과거 대화 회상 | ✅ hybrid 작동 (vec-daemon 필요); 대체 fallback = BM25-only + ⚠ vec-daemon down 표시 |
| bm25-memory | ~/.claude/hooks/bm25-memory.py |
UserPromptSubmit (--rich) |
G1 (의사결정) + G2-DOCS (docs/research/.md + CLAUDE/MEMORY) + G2-PREFETCH (codebase-memory-mcp) + G2-HOOKS (~/.claude/hooks/.py BM25 검색) — 단일 hook에 통합 | ✅ 프로덕션. pure BM25 (semantic rerank 없음). git-memory.py + g2-augment.py 대체 |
| memory-keyword-trigger | ~/.claude/hooks/memory-keyword-trigger.py |
UserPromptSubmit | 결정 키워드(제외/kill/중단/데드라인 등) 감지 → MEMORY.md 즉시 기록 유도 | ✅ 프로덕션. 저장은 Claude가 수행 |
| g2-fallback | ~/.claude/hooks/g2-fallback.py |
PostToolUse (Grep) | Grep 실패/빈약 결과 감지 → mcp__code-search__search_code 힌트 주입 |
✅ 프로덕션 |
| vec-daemon startup guard | ~/.claude/settings.json SessionStart |
SessionStart | vec-daemon 실 연결 probe → 죽어있으면 rm stale socket + 재시작 (2026-04-17 수정) | ✅ 프로덕션 |
은퇴한 / 아카이브 hooks (wired 아님, 실행 안 됨)
| 파일 | 상태 | 은퇴 근거 |
|---|---|---|
git-memory.py |
retired — bm25-memory.py에 의해 대체 (bm25-memory.py docstring 자체가 "Replaces git-memory.py"로 명시) |
recall 0.169 proactive → BM25 reactive로 성능 개선 |
g2-augment.py |
retired — bm25-memory.py G2-PREFETCH 블록에 흡수 |
동일 |
auto-index.py |
wired 아님 — settings.json에 없음. 코드는 존재하나 Claude Code가 invoke 안 함 | codebase-memory-mcp는 auto-index 없이 stale-on-demand로 관리 |
2026-04-17 세션 주의: 이번 세션의 iter 1-15 작업 대부분은
git-memory.py와auto-index.py에 대한 편집이었으나, 두 파일 모두 production에 wired 되어 있지 않음 → 해당 편집은 파일에 남아있지만 실제로 실행되지 않음. 진짜 production 영향이 있는 편집은chat-memory.py(iter 2 ⚠ 표시, iter 15 telemetry) 뿐.
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.
- 2d ago First seen · 246 lines · 3,854 tokens per session scan B b7468f3b56a9
CTX CLAUDE.md is an instructions file published in the GitHub repository pluto2060/CTX (9 stars, last pushed 3mo ago), licensed MIT. It adds 3,854 tokens to every session, about $0.0193 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
tdd-guard CLAUDE.md
Instructions for nizos/tdd-guard, covering tdd guard, project goal, development workflow, commit guidelines and project structure.
token-goat AGENTS.md
AGENTS.md instructions for DFKHelper/token-goat, covering agents.md, project, build, test, lint, the index must stay bounded and security boundaries.
token-goat copilot-instructions.md
Copilot instructions for DFKHelper/token-goat: Gate — before every file read, answer one question first: is there a token-goat command that returns just what I need? If yes, run it. A read tool invoked without answering the gate is a violation, not an oversight. The gate is per file: batched or parallel reads do not…
kglite CLAUDE.md
Instructions for kkollsga/kglite, covering kglite — claude code conventions, build & test, architecture, the boundary principle (wrappers vs core) — summary and in-memory is the core product.
code-context AGENTS.md
Instructions for infino-ai/code-context, covering code-context: notes for ai agents, project overview, repo map, build, test, gates and conventions.
klaudiush CLAUDE.md
Instructions for smykla-skalski/klaudiush, covering claude.md, project overview, commands, completion (shell completion scripts) and doctor (diagnose setup and configuration).