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/eunomia-bpf/agentsight/claude-mdgit clone --depth 1 https://github.com/eunomia-bpf/agentsightWrote 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/eunomia-bpf/agentsight/claude-md)<a href="https://agentmods.dev/instructions/eunomia-bpf/agentsight/claude-md"><img src="https://agentmods.dev/badge/instructions/eunomia-bpf/agentsight/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 5.1 | $0.02524 | $0.02524 |
| Opus 5 | $0.01262 | $0.01262 |
| Sonnet 5 | $0.00505 | $0.00505 |
| Haiku 4.5 | $0.00252 | $0.00252 |
Grade D, and why
agentsight 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 6d 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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
sudo ./agentsight record -- claude 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.
`sudo`, it falls back to the invoking user's `~/.kube/config` when How it starts
The opening of the file, as written. The whole thing — 177 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to coding agents working in this repository.
AGENTS.md is a symlink to this file so Claude Code and other agents share the
same repository instructions.
Overview
AgentSight is an eBPF-based observability framework for monitoring AI agent behavior through SSL/TLS traffic interception and process monitoring. It captures unencrypted request/response data at the kernel level without requiring any code changes to target applications.
Build & Test Commands
# Full build (eBPF + Rust collector + frontend)
make build
# Individual components
make build-bpf # eBPF C programs only
cd collector && cargo build --release # Rust collector only
cd frontend && npm install && npm run build # Frontend only
# Tests
cd bpf && make test # C unit + runtime tests
cd collector && cargo test # Rust tests
cd frontend && npm run lint # Frontend linting
# Run a single Rust test
cd collector && cargo test test_name
# Debug builds with AddressSanitizer
cd bpf && make debug
cd bpf && make sslsniff-debug
# Install system dependencies (Ubuntu/Debian)
make install
Running
# Live agent sessions
./agentsight top
# Launch and record a command
sudo ./agentsight record -- claude
# Record Claude Code (requires --binary-path for statically-linked BoringSSL)
sudo ./agentsight record -c claude --binary-path ~/.local/share/claude/versions/<version>
# Record Python AI tools
sudo ./agentsight record -c python
# Record with NVM Node.js (statically-linked OpenSSL)
sudo ./agentsight record -c node --binary-path ~/.nvm/versions/node/v20.0.0/bin/node
# Direct eBPF program usage
sudo ./bpf/sslsniff --binary-path <path>
sudo ./bpf/process -c python
# Web UI available at http://127.0.0.1:7395 when using record live capture.
# debug trace needs --server.
Documentation Hygiene
- Keep the README Quick Start stable. Do not update Quick Start unless the primary onboarding command or first-run flow changes.
- Put details about mode-specific behavior, persistence paths, storage formats, and operational caveats in Usage, FAQ, or dedicated docs sections instead of Quick Start.
- When changing user-facing CLI behavior, update the focused reference docs and examples that describe that behavior, but avoid broad README churn.
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.
- 6d ago First seen · 177 lines · 2,524 tokens per session scan D 4908d9c64c1a
agentsight CLAUDE.md is an instructions file published in the GitHub repository eunomia-bpf/agentsight (666 stars, last pushed yesterday), licensed MIT. It adds 2,524 tokens to every session, about $0.0126 per session on Opus 5. A static security scan graded it D with 2 findings (asks for root, reaches for credential files). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
hoop CLAUDE.md
Claude Code instructions for hoophq/hoop, covering claude.md, project overview, toolchain & prerequisites, architecture and module breakdown.
gini-agent AGENTS.md
AGENTS.md instructions for Open-Curiosity/gini-agent, covering gini agent instructions, shape, adrs, boundaries and branches.
rome CLAUDE.md
Claude Code instructions for rome-os/rome, covering rome, playbook and traps.
zapcode CLAUDE.md
Instructions for TheUncharted/zapcode, covering claude.md, git, read agents.md first, codebase orientation and how to add a new language feature.
MathModelAgent CLAUDE.md
Claude Code instructions for jihe520/MathModelAgent, covering claude.md, 项目概述, commands, 后端 and 安装依赖.
ralph CLAUDE.md
Claude Code instructions for SantanderAI/ralph, covering ralph — instrucciones del proyecto and paridad bash ↔ powershell (importante).