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 skills/wandererer/codebeacon/skillnpx skills add Wandererer/codebeacon --skill skillgit clone --depth 1 https://github.com/Wandererer/codebeaconWhat 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 | $0.00099 | $0.04401 |
| Opus 5 | $0.00049 | $0.02201 |
| Sonnet 5 | $0.00020 | $0.00880 |
| Haiku 4.5 | $0.00010 | $0.00440 |
Grade B, and why
codebeacon 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 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
| Codex CLI | `spawn_agent` (requires `multi_agent=true` in `~/.codex/config.toml`) | No model picker in subagent API — use the host's default | How it starts
The opening of the file, as written. The whole thing — 327 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/codebeacon
Scan source code with AST analysis → build a knowledge graph → generate a navigable wiki + CLAUDE.md context map ready for AI agents. After scan/sync, run AI-semantic enrichment automatically using whatever model the agent is currently running on (no hardcoded model, no extra API key).
Usage
/codebeacon # scan current directory + auto AI-semantic
/codebeacon <path> # scan specific path + auto AI-semantic
/codebeacon <path> --update # incremental scan + auto AI-semantic
/codebeacon <path> --wiki-only # wiki only (no scan, no semantic)
/codebeacon <path> --no-semantic # skip AI-semantic
/codebeacon sync # multi-project sync + auto AI-semantic
/codebeacon semantic-prepare [--dir D] # emit a fresh tasks file only (no scan)
/codebeacon semantic-apply [--dir D] # merge a results file written by the agent
/codebeacon serve <path> # start MCP server pointing at .codebeacon/
/codebeacon query <term> # search the graph
/codebeacon path <src> <tgt> # shortest path
/codebeacon knowledge [<path>] # scan .md notes (ADRs, meetings, retros, specs) → KNOWLEDGE.md
/codebeacon init [<path>] # interactive codebeacon.yaml
/codebeacon install # install/update this skill
/codebeacon upgrade # pip upgrade codebeacon + refresh this SKILL.md (then restart Claude Code)
If no path was given, use . (current directory). Do not ask the user for a path.
Step 0 — Constrained query expansion (for /codebeacon query flows)
When the user's question doesn't literally match graph labels (e.g. they ask in Korean about a Java codebase, or use a synonym), do NOT invent extra search terms from memory. Phantom expansion ("auth" → {passport, sso, saml, oauth, jwt, scim, …}) returns ungrounded results.
Instead, before running codebeacon query:
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 · 327 lines · 99 tokens per session scan B 2567e25ff225
codebeacon is a skill published in the GitHub repository Wandererer/codebeacon (2 stars, last pushed 1mo ago), licensed MIT. It adds 99 tokens to every session and 4,401 once invoked, about $0.0005 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 skills, from other repositories
ch015-va
Run CH015 vulnerability assessment workflows for source-code security architecture review, 8-dimension analysis, compliance checks, self-verification, evidence verification, and VA reporting.
ch015-verifier
Run CH015 adversarial verification workflows to independently validate security reports, recheck file-line evidence, find false positives, severity errors, missed findings, and dependencies.
ch015-feedback
Run CH015 security design review workflows for PRDs and specs to produce security opinions or review findings mapped to standards and regulatory requirements.
ch015-pentest
Run CH015 penetration testing workflows for attacker-minded source-code review, exploit scenarios, POC generation, dry-run or live verification, and security findings.
ch015
Run CH015 AI Security Firm workflows for security audits, vulnerability assessment, adversarial verification, pentesting, red team review, compliance checks, fix guidance, and report generation.
test-harness
Test codeweave-mcp tools against any project using TestHarness. Use this whenever verifying tool behavior, testing a code change, or running custom assertions — e.g. "test the harness", "run tests against X project", "does semanticsearch work correctly".