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/safedep/gryph/agents-mdgit clone --depth 1 https://github.com/safedep/gryphWrote 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/safedep/gryph/agents-md)<a href="https://agentmods.dev/instructions/safedep/gryph/agents-md"><img src="https://agentmods.dev/badge/instructions/safedep/gryph/agents-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 | $0.01030 | $0.01030 |
| Opus 5 | $0.00515 | $0.00515 |
| Sonnet 5 | $0.00206 | $0.00206 |
| Haiku 4.5 | $0.00103 | $0.00103 |
Grade B, and why
gryph AGENTS.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 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- Hooks configured in `~/.claude/settings.json` (per official docs) How it starts
The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Gryph
Build
make deps # Install dependencies
make generate # Generate ent code
make gryph # Build binary
make test # Run tests
make lint # Run linter (golangci-lint)
Always run make lint before submitting changes to catch issues early.
When modifying the Event model (core/events/event.go) or any payload types, run make generate-schema.
Architecture
core/ Domain models (events, sessions, audit, security) - most stable
config/ Viper-based configuration
storage/ SQLite + ent ORM
agent/ Adapter pattern (claudecode/, cursor/ and more)
cli/ Cobra commands as an App pattern
tui/ Output formatters (table, json, csv)
Key Entry Points
cmd/gryph/main.go- Entry pointcli/root.go- App struct, dependency injectionagent/adapter.go- Agent adapter interfacestorage/storage.go- Store interface
Notes
- SQLite driver:
modernc.org/sqlite(pure Go, usessqlitenotsqlite3) - Config paths: macOS
~/Library/Application Support/safedep/gryph/, Linux~/.config/safedep/gryph/. Overrides:GRYPH_CONFIG_DIR,GRYPH_DATA_DIR,GRYPH_CACHE_DIR - Storage layer fully implemented with ent ORM
- Self-audit logs tool actions (install, uninstall, config changes)
Hook Implementation
Claude Code
- Hooks configured in
~/.claude/settings.json(per official docs) - Hook types:
PreToolUse,PostToolUse,PostToolUseFailure,SessionStart,SessionEnd,Notification - Uses matcher pattern
"*"for tool hooks to capture all tools - Version detection via
claude -vcommand
Cursor
- Hooks configured in
~/.cursor/hooks.json - Hook types:
beforeSubmitPrompt,beforeShellExecution,beforeMCPExecution,beforeReadFile,afterFileEdit,stop
Session Tracking
- SessionID: Deterministic UUID derived from agent's session_id using
uuid.NewSHA1(uuid.NameSpaceOID, []byte(agentSessionID)) - AgentSessionID: Original session ID string from the agent, stored for correlation with agent's own storage (e.g., Claude Code transcripts)
- Sessions are created on first event and updated on subsequent events
- Session end detected from
SessionEnd(Claude Code) orstop(Cursor) hook types
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 · 92 lines · 1,030 tokens per session scan B 6f7189e48213
gryph AGENTS.md is an instructions file published in the GitHub repository safedep/gryph (161 stars, last pushed 4d ago), licensed Apache-2.0. It adds 1,030 tokens to every session, about $0.0052 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 2026-08-30.
Other instructions, from other repositories
coral CLAUDE.md
Instructions for cdknorow/coral, covering claude.md - coral go, mission, testing, go unit tests and legacy parity tools (historical reference).
research-cli AGENTS.md
Instructions for rmedranollamas/research-cli, covering agent notes, architecture, gemini cli extension, security notes and development.
gemini-cli-gopls GEMINI.md
Instructions for Menghuan1918/gemini-cli-gopls, covering the gopls mcp server, init a go workspace, detecting a go workspace, go programming workflows and read workflow.
azure-sdk-for-go go-code.instructions.md
Instructions for Azure/azure-sdk-for-go: All code should follow the guidelines from the Azure Go SDK Guidelines. This document is a summary of the most important guidelines to follow when contributing to the Azure Go SDK.
aider-desk AGENTS.md
AGENTS.md instructions for hotovo/aider-desk, covering agents.md, common commands, development, type checking and linting and formatting.
chatgpt-cli CLAUDE.md
Claude Code instructions for kardolus/chatgpt-cli, covering chatgpt-cli — release runbook, prerequisites, 1. cut the release, 2. publish the github release + binaries and 3. update the homebrew tap.