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/odradekai/bundles-forge/agents-mdgit clone --depth 1 https://github.com/OdradekAI/bundles-forgeWrote 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/odradekai/bundles-forge/agents-md)<a href="https://agentmods.dev/instructions/odradekai/bundles-forge/agents-md"><img src="https://agentmods.dev/badge/instructions/odradekai/bundles-forge/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 | $0.01555 | $0.01555 |
| Opus 5 | $0.00777 | $0.00777 |
| Sonnet 5 | $0.00311 | $0.00311 |
| Haiku 4.5 | $0.00155 | $0.00155 |
Grade A, and why
bundles-forge AGENTS.md scanned grade A 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- **NEVER** add `eval()`, `child_process`, network calls (`curl`/`wget`) in hook scripts or plugin code Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
- **NEVER** add `eval()`, `child_process`, network calls (`curl`/`wget`) in hook scripts or plugin code How it starts
The opening of the file, as written. The whole thing — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PROJECT KNOWLEDGE BASE
Generated: 2026-04-17 Commit: 8c2e762 Branch: main
OVERVIEW
Bundle-plugin engineering toolkit (Python 3.9+, Node.js packaging). 8 skills in hub-and-spoke architecture: 3 orchestrators dispatch 4 executors. Supports 6 platforms. Dogfoods its own patterns.
STRUCTURE
bundles-forge/
├── bin/ # CLI dispatcher → routes subcommands to skills/*/scripts/
├── skills/ # 8 skills, each with SKILL.md + optional references/
│ ├── auditing/ # scripts/ (7 Python scripts), references/ (11 files)
│ └── scaffolding/ # assets/ (6 platform templates + root/hooks/scripts)
├── agents/ # 3 read-only subagents: inspector, auditor, evaluator
├── hooks/ # session-start (Bash) + run-hook.cmd (polyglot) + openclaw-bootstrap/
├── docs/ # 10 guide pairs (EN + .zh.md), checked by audit-docs D7-D9
├── tests/ # 6 suites via run_all.py; fixtures/ + prompts/ per skill
├── examples/ # Worked audit report examples
├── .github/workflows/ # validate-plugin.yml (Python 3.9+3.12 matrix)
└── package.json # v1.7.8 — version synced via .version-bump.json
WHERE TO LOOK
| Task | Location | Notes |
|---|---|---|
| Add audit check | skills/auditing/references/audit-checks.json |
Registry of all S1-PC7 checks; scripts read this |
| Fix audit script | skills/auditing/scripts/ |
Shared _cli.py for argparse/exit codes; _graph.py for workflow DAG |
| Add platform support | skills/scaffolding/assets/platforms/<name>/ |
Template files per platform (plugin.json, hooks.json, AGENTS.md, etc.) |
| Version management | skills/releasing/scripts/bump_version.py |
Driven by .version-bump.json; never edit versions manually |
| Session bootstrap | hooks/session-start + hooks/run-hook.cmd |
3-way platform detection (CURSOR_PLUGIN_ROOT, CLAUDE_PLUGIN_ROOT, COPILOT_CLI/fallback) |
| Security patterns | skills/auditing/references/security-checklist.md |
7 surfaces: skill content, hook scripts, HTTP hooks, OpenCode, agents, bundled scripts, MCP |
| Test a skill | tests/prompts/<skill-name>.yml |
YAML prompt samples for should-trigger/should-not-trigger |
| Add a test fixture | tests/fixtures/<scenario>/ |
Simulate plugin states (artifact-mismatch, circular-deps, etc.) |
| Documentation sync | docs/*.md ↔ docs/*.zh.md |
Every guide has a Chinese translation; D7-D9 checks consistency |
| CLI subcommand | bin/bundles-forge |
Dispatcher; add new subcommand to COMMANDS dict |
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 · 100 lines · 1,555 tokens per session scan A 33e2690dd93c
bundles-forge AGENTS.md is an instructions file published in the GitHub repository OdradekAI/bundles-forge (230 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 1,555 tokens to every session, about $0.0078 per session on Opus 5. A static security scan graded it A with 2 findings (makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
uikit-expert-skill AGENTS.md
Instructions for zinxj/uikit-expert-skill, covering agent guidelines for uikit expert skill, core principles, 1. uikit focus only, 2. no code formatting or linting and 3. no architectural opinions.
uikit-expert-skill CLAUDE.md
Instructions for zinxj/uikit-expert-skill, a project described as: Provide accurate, efficient, and modern UIKit Swift code using an AI agent skill focused on best practices and performance optimization.
ai CLAUDE.md
Claude Code instructions for formio/ai, covering claude.md, project overview, repository info, skills library and iterating on skills.
Looptimal CLAUDE.md
Instructions for Renn-Labs/Looptimal, covering looptimal — repo guidance for ai agents, build journal — capture the concepts, not the changelog and — .
community CLAUDE.md
Claude Code instructions for nowledge-co/community, covering nowledge community — agent guidelines, registry, universal install contract — skill.md, behavioral guidance and plugin development.
superpowers-graph CLAUDE.md
Claude Code instructions for RonMizrahi/superpowers-graph, covering superpowers-graph — repository guide, sources of truth, changing the graph — the eval ships with the change and rules that are not obvious from the code.