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/slxca/opencontext/agents-mdgit clone --depth 1 https://github.com/slxca/opencontextWrote 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/slxca/opencontext/agents-md)<a href="https://agentmods.dev/instructions/slxca/opencontext/agents-md"><img src="https://agentmods.dev/badge/instructions/slxca/opencontext/agents-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 5.1 | $0.01146 | $0.01146 |
| Opus 5 | $0.00573 | $0.00573 |
| Sonnet 5 | $0.00229 | $0.00229 |
| Haiku 4.5 | $0.00115 | $0.00115 |
Grade A, and why
opencontext 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 yesterday.
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 — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CRITICAL MEMORY SYSTEM: OpenContext Protocol
You have access to persistent project context via OpenContext MCP tools (.opencontext/). Adherence to this protocol is mandatory for consistency across sessions.
1. Mandatory Pre-Flight: Inspect Context
Before editing, refactoring, or generating code that affects:
- Architecture, module boundaries, or dependency graph
- Database schemas, migrations, or models
- API contracts, route handlers, or DTOs
- Authentication, authorization, or security rules
Action Required: Execute read_context FIRST to check existing conventions, past design decisions, and project rules. Do NOT assume architectural patterns without reading relevant context topics.
2. Post-Implementation: Persist Changes
Immediately after introducing, altering, or deprecating conventions, run the corresponding OpenContext tool:
- Save/Update (
save_context):- When: You add or change an architectural pattern, schema, shared utility rule, or tech stack decision.
- Format: Concise, topic-scoped Markdown summary.
- Lifecycle Frontmatter: When updating an existing convention, you MUST include YAML frontmatter:
--- status: active | superseded | deprecated supersedes: previous_topic_slug ---
- Deprecate vs. Delete (
delete_context):- Use
status: deprecatedorstatus: supersededin YAML frontmatter if historical context is valuable. - Call
delete_contextONLY when a topic is fully obsolete or entirely removed from the codebase.
- Use
3. Execution Rules
- Never skip: Do not bypass
read_contexton multi-file or cross-cutting tasks to save tokens. - Fail-safe: If unsure whether an architectural decision exists, invoke
read_contextbefore taking action.
Repository
pnpm monorepo ([email protected]). Two packages:
packages/opencontext— theopencontext-mcpMCP server (published npm package, the real product).apps/web— private Next.js marketing/docs site (@opencontext/web).
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.
- yesterday First seen · 75 lines · 1,146 tokens per session scan A f284945034c4
opencontext AGENTS.md is an instructions file published in the GitHub repository slxca/opencontext (11 stars, last pushed 2d ago), licensed MIT. It adds 1,146 tokens to every session, about $0.0057 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 2026-09-04.
Other instructions, from other repositories
capsule AGENTS.md
Instructions for dawitlabs/capsule, a project described as: Markdown context packs that help coding agents reduce repeated repository-discovery tokens.
ratel CLAUDE.md
Claude Code instructions for ratel-ai/ratel, a project described as: Context engineering for AI agents. 80% fewer tokens. Fix tool overload. Skills and memory with in-process BM25 and semantic retrieval. Progressive Disclosure. No vector DB.
ctx4-ai AGENTS.md
AGENTS.md instructions for dubzdubz/ctx4-ai, covering ctx4-ai - agent context, quick reference, project structure, architecture and key design decisions.
mcp-file-context-server CLAUDE.md
Instructions for bsmi021/mcp-file-context-server, covering claude.md, development commands, build and run, environment configuration and architecture overview.
capsule CLAUDE.md
Instructions for dawitlabs/capsule, a project described as: Markdown context packs that help coding agents reduce repeated repository-discovery tokens.
ctx4-ai CLAUDE.md
Claude Code instructions for dubzdubz/ctx4-ai, a project described as: Your portable context layer for Claude & ChatGPT.