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/ducksss/lifehack-2026/agents-mdgit clone --depth 1 https://github.com/Ducksss/LifeHack-2026What 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 | $0.01786 | $0.01786 |
| Opus 5 | $0.00893 | $0.00893 |
| Sonnet 5 | $0.00357 | $0.00357 |
| Haiku 4.5 | $0.00179 | $0.00179 |
Grade A, and why
LifeHack-2026 AGENTS.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 3d 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 — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Woven agent instructions
This repository is the LifeHack 2026 submission workspace. The user-facing
product is Woven; LifeHack-2026 is only the repository name.
Canonical context
Before changing the project, inspect git status, preserve unrelated user work,
and read these files in order:
docs/HANDOVER.md— current decisions, implementation status, external state, archived branches, and next workdocs/PRD.md— binding implemented product contract and acceptance criteriadocs/architecture.md— tool contracts, state machine, and trust boundariesREADME.md— public story, setup, screenshots, and working demo instructions
Then route by task:
- Pitch, judging, recording, or demo: read
script.md, thendocs/DEVPOST_SUBMISSION.mdanddocs/BRAND_GUIDE.md. - Plugin, ChatGPT, Codex, or local installation: read
docs/INSTALLATION.md,.codex-plugin/plugin.json,.mcp.json, and.agents/plugins/marketplace.json. - Branding, slides, or gallery assets: read
docs/BRAND_GUIDE.md; the current judge deck isdocs/Woven-Hackathon-Pitch.pptx.
When documentation conflicts, implemented behavior and tests win. Use this
precedence: code/tests and docs/PRD.md → docs/architecture.md →
docs/HANDOVER.md → public/pitch material. A storyboard never makes a planned
feature real.
Current snapshot
- The working product is one Node.js service with a React MCP App, HTTP/stdio MCP
transports, a bounded LangGraph.js mission-orchestration layer, deterministic
commerce verification,
/demo,/webmcp,/identity,/merchant,/install, SQLite state, and a simulated Visa adapter. /webmcpexposes seven top-level browser-native site tools over the same server router and visible Choice Center; identity and purchase remain human-only.- The canonical rainy-weekend camping request produces complete, weather-compatible, one-merchant carts and an exact, expiring confirmation.
- Non-camping requests route through a credential-dormant open-world POC that
interprets a validated
MissionSpec, discovers connected offers and cited web research, composes and verifies carts, retries at most once, and fails closed when OpenAI is unavailable. Web research is never checkout evidence. - The connector-style demo identity check is implemented and enforced before checkout with a short-lived session, state, PKCE, an allowlisted callback, and a single-use code. It is simulated and never a Visa login or KYC claim.
- The judge deck includes the working demo identity handoff between cart choice and exact checkout review.
- Public HTTPS deployment and the demo video are complete. A shareable ChatGPT connection, real merchant/Visa integrations, and final Devpost publication remain outstanding.
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.
- 3d ago First seen · 149 lines · 1,786 tokens per session scan A 6898e5e3383c
LifeHack-2026 AGENTS.md is an instructions file published in the GitHub repository Ducksss/LifeHack-2026 (5 stars, last pushed 3d ago), licensed MIT. It adds 1,786 tokens to every session, about $0.0089 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 2026-08-31.
Other instructions, from other repositories
mnemopay-sdk CLAUDE.md
Instructions for mnemopay/mnemopay-sdk, covering mnemopay sdk, quick start, architecture, two modes and key flows.
optifeed-radar CLAUDE.md
Instructions for optifeed/optifeed-radar, covering claude.md - optifeed radar, commands, stack, layout and hard rules (violations = rejected pr).
consult CLAUDE.md
Claude Code instructions for kreek/consult, covering claude.md and pi self-improvement source access.
lianhuanhua-skills AGENTS.md
Instructions for littlewindy123/lianhuanhua-skills, covering agents.md, project goal, architecture rules, commands and before committing.
ai-plugin GEMINI.md
Gemini CLI instructions for PostHog/ai-plugin, covering posthog extension, available tool categories and guidelines.
agent-skills GEMINI.md
Gemini CLI instructions for motherduckdb/agent-skills, covering motherduck skills for gemini cli, default routing, skill catalog, utility and workflow.