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/blisspixel/primr/agents-mdgit clone --depth 1 https://github.com/blisspixel/primrWhat 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 | $0.05413 | $0.05413 |
| Opus 5 | $0.02707 | $0.02707 |
| Sonnet 5 | $0.01083 | $0.01083 |
| Haiku 4.5 | $0.00541 | $0.00541 |
Grade A, and why
primr 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 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.
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 — 287 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md - primr
This file is portable agent guidance in the agents.md format. AI coding tools that auto-detect AGENTS.md (Kiro, OpenAI Codex, Jules, Aider, others) load it automatically. Tools that don't (Claude Code, Cursor, Windsurf) can reference or include it from their own rules / skill files.
Operating primr vs. building primr. This file is for agents that operate the primr CLI/MCP to do research. If you are an agent (or human) here to change primr's source code, read
CLAUDE.md- the development contract - not this file.
If you are an AI agent reading this in a primr-aware project: this is how to route and operate Primr without making the user learn its internal commands.
Primr has two serious research paths. Primr Zero combines deterministic Primr collection with the current agent host's research and reasoning. The provider-backed pipeline adds Primr-managed model synthesis, cross-validation, strategy generation, and rendering after an estimate and explicit spend approval. Both aim for useful strategic artifacts; the host-powered path is not a renamed scrape mode or a deliberately worse application.
Agent-host default
When a request is addressed to an agent host, keep the public invocation simple:
primr "Company" https://company.example
Treat that bare company-and-URL request, "run Primr," and "build a full
dossier" as Primr Zero by default. Use the primr-zero skill and run
primr prep internally when a shell is available. Do not make the user choose
between prep, scrape, and full, and do not ask for spend approval for the
zero-model-call collection. Disclose its public network activity, then
continue.
Route to the provider-backed pipeline only when the user explicitly asks for
paid, metered, provider-backed, or premium Primr; supplies a dollar budget;
asks to use provider API keys; or provides provider-only CLI modifiers such as
--premium, --mode, --platform, --strategy-type, or
--no-ai-strategy. Configured API keys are capability, not consent to spend.
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 · 287 lines · 5,413 tokens per session scan A 8711accb0f7c
primr AGENTS.md is an instructions file published in the GitHub repository blisspixel/primr (3 stars, last pushed 2d ago), licensed Apache-2.0. It adds 5,413 tokens to every session, about $0.0271 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
designing-real-world-ai-agents-workshop CLAUDE.md
Instructions for iusztinpaul/designing-real-world-ai-agents-workshop, covering project, project structure, tech stack, access documentation and running qa.
opencolab AGENTS.md
Instructions for jimezsa/opencolab, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.
epistemedia AGENTS.md
Instructions for yoheinakajima/epistemedia, covering agents.md — epistemedia control contract, governing rule, mandatory orientation, authority map and epistemic rules.
cortex-app CLAUDE.md
Instructions for mocaOS/cortex-app, covering claude.md, navigation map, file-path routing, priority and meta: maintaining these docs.
rockie-claude CLAUDE.md
Instructions for Rockielab/rockie-claude, covering how to work in rockie-claude, first reads (after the workspace docs/ reads), ⚠️ this repo is a local overlay, not the canonical source, process for changes here and skills you'll reach for.
autoresearch-paper AGENTS.md
Instructions for WdBlink/autoresearch-paper: Project context & docs live in ./ctx — managed by the ctx skills; apply their consistency rules when editing ctx docs.