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/kelp/webdown/claude-mdgit clone --depth 1 https://github.com/kelp/webdownWhat 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 | $0.01657 | $0.01657 |
| Opus 5 | $0.00829 | $0.00829 |
| Sonnet 5 | $0.00331 | $0.00331 |
| Haiku 4.5 | $0.00166 | $0.00166 |
Grade B, and why
webdown CLAUDE.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 2d 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.
Unrestricted tool accessmediumExcessive agency
A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.
- Use `./scripts/run.sh <command>` to run any command in the correct environment How it starts
The opening of the file, as written. The whole thing — 155 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Webdown Development Guide
Working with Claude
- No hallucinations: If Claude is uncertain about any answer or how to do something, it should explicitly say so rather than guessing or making up information.
- Clear uncertainty: When encountering ambiguous requirements or unknown aspects of the codebase, Claude should openly acknowledge uncertainty and ask for clarification.
- Providing alternatives: When unsure about a specific solution, Claude should present multiple options along with their potential trade-offs.
Version Management (Semantic Versioning)
We follow Semantic Versioning for version numbering (MAJOR.MINOR.PATCH):
- Patch version (0.4.1 → 0.4.2): Documentation updates, bug fixes, and other changes that don't modify the API
- Minor version (0.4.2 → 0.5.0): New features or enhancements that maintain backward compatibility
- Major version (0.9.0 → 1.0.0): Breaking changes, significant rewrites, or major feature overhauls
When making changes, choose the appropriate version increment based on the impact of your changes.
Modern Python Development Setup (Python 3.10+)
Build & Testing Commands
- Install dependencies:
poetry install - Run all tests:
poetry run pytestormake test - Run specific test:
poetry run pytest webdown/tests/test_file.py::TestClass::test_function - Run with coverage:
poetry run pytest --cov=webdownormake test-coverage - Run integration tests:
poetry run pytest --integrationormake integration-test - Type checking:
poetry run mypy webdownormake type-check - Linting:
poetry run flake8 webdownormake lint - Formatting:
poetry run black webdown && poetry run isort webdownormake format - All checks:
make all-checks - Building package:
poetry buildormake build
Path-Independent Environment Management
- We use a path-independent approach that works regardless of where the repo is cloned
- Poetry is configured to create virtualenvs in-project:
poetry config virtualenvs.in-project true - Use
./scripts/run.sh <command>to run any command in the correct environment - Examples:
- Run tests:
./scripts/run.sh pytest - Run pre-commit:
./scripts/run.sh pre-commit run - Run any other command:
./scripts/run.sh python -m webdown.cli
- Run tests:
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.
- 2d ago First seen · 155 lines · 1,657 tokens per session scan B ebfda8c003bf
webdown CLAUDE.md is an instructions file published in the GitHub repository kelp/webdown (6 stars, last pushed 16d ago), licensed MIT. It adds 1,657 tokens to every session, about $0.0083 per session on Opus 5. A static security scan graded it B with 1 finding (unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
rdrr CLAUDE.md
Instructions for fkonovalov/rdrr, covering claude.md, what is this project?, project structure, common commands and code conventions.
docpull AGENTS.md
Instructions for raintree-technology/docpull: DocPull is the open-source context dependency manager for AI agents. Its core contract is local-first, browser-free by default, reproducible, cited, and budget guarded.
agentscout AGENTS.md
AGENTS.md instructions for agentx402-ai/agentscout, covering agents.md — agent/contributor guide for this repo, what this repo is, commands, conventions and money-safety invariants (do not weaken).
agentscout CLAUDE.md
Claude Code instructions for agentx402-ai/agentscout: The shared agent/contributor guide lives in AGENTS.md (cross-tool standard) — commands, conventions, money-safety invariants, release process. Read it first.
markdown_web_browser AGENTS.md
AGENTS.md instructions for Dicklesworthstone/markdown_web_browser, covering agents.md — markdownwebbrowser, rule 0 - the fundamental override prerogative, rule number 1: no file deletion, git branch: only use main, never master and toolchain: python & uv.
firecrawl CLAUDE.md
Claude Code instructions for firecrawl/firecrawl: Firecrawl is a web scraper API. The directory you have access to is a monorepo.