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/omril321/claude-code-reflector/claude-mdgit clone --depth 1 https://github.com/omril321/claude-code-reflectorWrote 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/omril321/claude-code-reflector/claude-md)<a href="https://agentmods.dev/instructions/omril321/claude-code-reflector/claude-md"><img src="https://agentmods.dev/badge/instructions/omril321/claude-code-reflector/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 | $0.02217 | $0.02217 |
| Opus 5 | $0.01108 | $0.01108 |
| Sonnet 5 | $0.00443 | $0.00443 |
| Haiku 4.5 | $0.00222 | $0.00222 |
Grade D, and why
claude-code-reflector CLAUDE.md scanned grade D with 2 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 4d 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.
Reaches for credential fileshighPrivilege escalation
SSH keys, cloud credentials, git-credentials, .npmrc, /etc/shadow: reading these is how a config file becomes a credential leak.
- `AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` - or use `~/.aws/credentials` Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
│ ├── settings-reader.ts # Read ~/.claude/settings.json, pattern matching How it starts
The opening of the file, as written. The whole thing — 143 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Code Reflector
Analyzes Claude Code sessions to detect missing CLAUDE.md rules, unused skills, and skill corrections. Uses a two-tier architecture: Haiku for broad candidate detection, Sonnet for deep verification with full conversation context.
Sibling project to ~/private/code-claude-auto-learn (which captures what the user explicitly teaches). This project detects what's missing or broken that the user may not have noticed.
Commands
| Command | Description |
|---|---|
yarn pipeline |
Full pipeline: scan → verify → report |
yarn pipeline -- --limit 50 |
Full pipeline on up to 50 sessions |
yarn pipeline -- --concurrency 3 |
Control parallel API calls (default: 5) |
yarn scan |
Run Tier 1 analysis (incremental, skips processed sessions) |
yarn scan:dry-run |
List matching sessions without API calls |
yarn scan -- --limit 5 |
Process at most 5 sessions |
yarn scan -- --session <id> |
Process a single session |
yarn scan -- --all |
Re-process everything, ignore state |
yarn verify |
Deep-verify latest Tier 1 findings with Sonnet |
yarn verify -- --model sonnet |
Verify with specific model |
yarn verify -- --dry-run |
Show what would be verified |
yarn permissions |
Analyze sessions and suggest missing permission configs |
yarn permissions -- --limit 50 |
Analyze up to 50 sessions |
yarn report |
View latest Tier 1 report (human-readable) |
yarn report -- --verified |
View latest verified report |
yarn report -- --permissions |
View latest permissions report |
yarn report -- --json |
View latest report as raw JSON |
yarn reset |
Clear processing state |
yarn build |
Compile TypeScript |
Architecture
src/
├── index.ts # CLI entry (commander)
├── types/
│ ├── session.ts # SessionIndexEntry, CondensedSession, RawJSONLEntry
│ ├── findings.ts # Tier1Flag, Tier2Verdict, reports
│ ├── permissions.ts # ToolUseRecord, PermissionPattern, PermissionReport
│ └── state.ts # ReflectorState, ProcessedSessionRecord
├── scanner/
│ ├── index-reader.ts # Session discovery (index + JSONL fallback), filter entries
│ ├── session-parser.ts # Stream JSONL → condensed conversation text
│ └── skill-catalog.ts # Load SKILL.md frontmatter + CLAUDE.md content
├── analyzer/
│ ├── anthropic-client.ts # Vertex/Bedrock SDK wrapper with retry + model config
│ ├── tier1.ts # Tier 1 (Haiku) broad analysis
│ ├── tier2.ts # Tier 2 (Sonnet) deep verification
│ ├── prompts.ts # Tier 1 system/user prompt templates
│ └── tier2-prompts.ts # Tier 2 verification prompt templates
├── permissions/
│ ├── extractor.ts # Stream JSONL → ToolUseRecord[] with approval tracking
│ ├── settings-reader.ts # Read ~/.claude/settings.json, pattern matching
│ ├── analyzer.ts # Bash normalization, aggregation, scope, overlap detection, annotations
│ ├── safety.ts # Blocklist + LLM safety assessment
│ └── reporter.ts # Permission report output + console summary
├── state/
│ └── manager.ts # Atomic JSON state persistence
└── reporter/
└── index.ts # JSON report output + console summary
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.
- 4d ago First seen · 143 lines · 2,217 tokens per session scan D 5b11e9a26769
claude-code-reflector CLAUDE.md is an instructions file published in the GitHub repository omril321/claude-code-reflector (2 stars, last pushed 5mo ago), licensed MIT. It adds 2,217 tokens to every session, about $0.0111 per session on Opus 5. A static security scan graded it D with 2 findings (reaches for credential files, 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
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.