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/eph5xx/tweakidea/claude-mdgit clone --depth 1 https://github.com/eph5xx/tweakideaWrote 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/eph5xx/tweakidea/claude-md)<a href="https://agentmods.dev/instructions/eph5xx/tweakidea/claude-md"><img src="https://agentmods.dev/badge/instructions/eph5xx/tweakidea/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 | $0.00840 | $0.00840 |
| Opus 5 | $0.00420 | $0.00420 |
| Sonnet 5 | $0.00168 | $0.00168 |
| Haiku 4.5 | $0.00084 | $0.00084 |
Grade A, and why
tweakidea CLAUDE.md scanned grade A with 0 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project
TweakIdea
A Claude Code skillset (slash commands + subagent definitions) that helps founders evaluate startup problems and discover product opportunities. Eight commands: /tweak:browse-hn searches HN via Algolia and ranks candidate threads to feed into /tweak:analyze-hn-post; /tweak:analyze-hn-post analyzes a single Hacker News discussion to identify technology shifts and surface product ideas; /tweak:evaluate runs 14 independent subagents to produce a weighted scorecard with assumption tracking; /tweak:improve reads a completed run and generates three concrete idea tweaks (small reframe, medium reshape, big reimagine) grounded in scoring rubrics; /tweak:diff compares two evaluation runs side-by-side showing score deltas, potential shifts, and per-dimension movers; /tweak:list and /tweak:show browse and open artifacts accumulated under ~/.tweakidea/; /tweak:share uploads a run's HTML report to a secret GitHub gist and returns a shareable rendered preview URL.
Core Value: Help founders make better decisions — evaluate whether a problem is worth solving, or discover what problems are emerging from technology shifts.
Constraints
- Platform: Claude Code CLI — slash commands + agent definitions only. Exceptions:
analyze-hn-postandbrowse-hnboth shell out touv+ Python (hnparse.pyandhnsearch.py) for HN fetching - Evaluation model: All 14 dimensions from EVALUATION.md must be covered, no skipping
- Independence: Each subagent evaluates its dimension without seeing other dimensions' results
- Clean context: Each evaluation run should be independent — no cross-contamination between runs
- JSON contracts: All pipeline artifacts are typed JSON validated against schemas in
schemas/. The orchestrator does zero text processing on agent output — all aggregation is inscripts/compute.py
Key Decisions
- Founder data location:
~/.tweakidea/FOUNDER.md - Evaluator model: Sonnet (cost/speed balance for 14 parallel agents)
- Narrative model: Opus (synthesis quality for cross-dimensional prose)
- Researcher model: Sonnet (web research in Prepare stage)
- Context isolation: FOUNDER.md passed ONLY to founder-market-fit evaluator; other 13 evaluators get EVALUATION_CONTEXT only
- HN data location:
~/.tweakidea/hn/hn-{id}/ - HN analysis model: Inline (no subagent — runs in command context)
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 · 50 lines · 840 tokens per session scan A 0f56020762d7
tweakidea CLAUDE.md is an instructions file published in the GitHub repository eph5xx/tweakidea (48 stars, last pushed 4mo ago), licensed MIT. It adds 840 tokens to every session, about $0.0042 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
social-media-research-skills AGENTS.md
AGENTS.md instructions for ScrapeCreators/social-media-research-skills, covering social media research skills repo, authoring guidelines and validation.
social-media-research-skills CLAUDE.md
Claude Code instructions for ScrapeCreators/social-media-research-skills: Use these skills as practical social media research workflows.
web-search-mcp CLAUDE.md
Claude Code instructions for sydasif/web-search-mcp, covering claude.md, common commands, development, testing and architecture.
tc AGENTS.md
AGENTS.md instructions for Leobai03/tc, covering tc 项目维护规则, 产品真源, 修改规则 and 发布前检查.
app-store-operator CLAUDE.md
Instructions for meyusufdemirci/app-store-operator, covering app store operator — mcp server, codebase, server wiring, tool module contract and prompts.
forbotsake AGENTS.md
Instructions for forbotsake/forbotsake, covering forbotsake, skills, getting started, key files (created by skills) and installation.