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/franrom/pupila/agents-mdgit clone --depth 1 https://github.com/FranRom/pupilaWhat 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 | $0.09832 | $0.09832 |
| Opus 5 | $0.04916 | $0.04916 |
| Sonnet 5 | $0.01966 | $0.01966 |
| Haiku 4.5 | $0.00983 | $0.00983 |
Grade A, and why
pupila AGENTS.md scanned grade A 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 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sI "https://api.ashbyhq.com/posting-api/job-board/<slug>?includeCompensation=true" How it starts
The opening of the file, as written. The whole thing — 403 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for future Codex sessions working in this repo.
Overview
pupila is a config-driven, forkable daily job aggregator. It fetches listings from 13 public sources (3 ATS APIs — Ashby, Greenhouse, Lever — plus RSS feeds, JSON job boards, Hacker News, HTML scrapers, an Aave Next.js scraper, and ashby-private — a config-driven fetcher for orgs hosted on Ashby with the public posting-API disabled), normalizes them, applies hard exclusion filters, computes a per-job fitScore, deduplicates, and writes data/jobs.json, an RSS feed at data/feed.xml, and an auto-regenerated JOBS.md table. The hand-written README.md is the project doc and is not rewritten by the pipeline. No external services. No DB. Output lives in this repo.
The repo ships a neutral template in config/profile.json. After onboarding (CV upload → brief generation), /api/profile-generate shells out to the local LLM CLI to fill in the personal keyword lists + weights based on the brief. Re-runnable from Settings → Scoring profile → Regenerate. config/slugs.json ships with the full ~50-company tier-S list (all public ATS URLs — non-personal data, edit by hand to add/remove companies).
First-run UX: a forker generates their config/candidate-brief.md by running pnpm run setup-brief --file ~/cv.pdf (or via the UI's Profile tab → drop a PDF/DOCX/MD CV). That CLI shells out to whichever local LLM CLI is installed (Codex, codex, gemini, opencode — auto-detected, override via PUPILA_LLM=<provider>).
Stack
- Node 22 LTS, ESM, TypeScript 5.9 (NodeNext)
- Biome 2.4 (lint + format, single config in
biome.json) - pnpm 10
- Vitest 3 (tests in
tests/, run viapnpm test— 120 cases) - simple-git-hooks (pre-commit
lint && typecheck) - Single runtime dep:
fast-xml-parser. Nativefetchonly.
Run locally
pnpm install # also installs the pre-commit hook
pnpm run dev # tsx, no build step (this is what the launchd/cron aggregate agent runs)
pnpm start # built output: requires pnpm run build first
pnpm run typecheck # tsc --noEmit on src/, then on src/+tests/ via tsconfig.test.json
pnpm run lint # biome check
pnpm run lint:fix # biome check --write
pnpm test # vitest run (120 unit tests)
pnpm run test:watch # vitest watch mode
pnpm run ui # local-only UI: Vite dev server on http://127.0.0.1:5173, reads data/jobs.json
pnpm run ai-review # local-only: shells out to `Codex -p` to write data/ai-reviews.json
pnpm run ai-review --top=50 # raise the per-run cap (default 20 highest-fitScore unreviewed)
pnpm run ai-review --force # re-review entries that already exist
pnpm run ai-review --ids=a,b # review specific job ids only
pnpm run setup-brief --file ~/cv.pdf # generate config/candidate-brief.md from a CV (PDF/DOCX/MD/TXT)
pnpm run daily # convenience: pnpm run dev && pnpm run ai-review (morning routine)
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 · 403 lines · 9,832 tokens per session scan A 90ad46ce1b81
pupila AGENTS.md is an instructions file published in the GitHub repository FranRom/pupila (35 stars, last pushed 2mo ago), licensed MIT. It adds 9,832 tokens to every session, about $0.0492 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
autotask-mcp dev_workflow.instructions.md
Guide for using Taskmaster to manage task-driven development workflows.
ai-workflow AGENTS.md
Instructions for cunhaax/ai-workflow, covering ai workflow template, rules — non-negotiable, project overview, commands and architecture.
trackly-cli CLAUDE.md
Claude Code instructions for trackly-app/trackly-cli, covering trackly-cli, tech stack, backend production source of truth, directory structure and key commands.
foggy-data-mcp-bridge-python CLAUDE.md
Instructions for foggy-projects/foggy-data-mcp-bridge-python, covering foggy data mcp bridge — python, 快速启动, 安装依赖, 运行测试 and 启动 mcp 服务（连接 docker mysql）.
McpServer AGENTS.md
AGENTS.md instructions for sharpninja/McpServer, covering agent instructions, session start, rules, byrd test gate and where things live.
AmbyKit CLAUDE.md
Instructions for ambystechcom/AmbyKit, covering claude code — notes for the ambykit repo and claude-specific.