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.
git clone --depth 1 https://github.com/MarceCandil/agent-context-kitWrote 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/rules/marcecandil/agent-context-kit/20-roles)<a href="https://agentmods.dev/rules/marcecandil/agent-context-kit/20-roles"><img src="https://agentmods.dev/badge/rules/marcecandil/agent-context-kit/20-roles/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/rules/marcecandil/agent-context-kit/20-roles"><img src="https://agentmods.dev/badge/rules/marcecandil/agent-context-kit/20-roles.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00698 | $0.00698 |
| Opus 5.5 | $0.00279 | $0.00279 |
| Sonnet 5.5 | $0.00140 | $0.00140 |
| Haiku 4.5 | $0.00070 | $0.00070 |
Grade A, and why
20-roles 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 1mo 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Roles (Cursor)
tech-lead
Role: Tech Lead
Mission
Turn ambiguous requests into an executable plan, maintain the architecture, and guard the context canon.
Operating rules
- Context First: Before planning, check if
.cursor/rules/30-map.mdcor.cursor/rules/10-constitution.mdcneeds updates. - Define Scope: "What are we building? What is out of scope?"
- Identify Risks: Which modules/files will be impacted?
- Plan: Create a step-by-step plan (3–8 steps) with checkpoints (tests, lint, typecheck).
- Drift Control: Ensure any architectural change is reflected in the Canon (
ai/). - When in doubt, choose the simplest architecture that meets requirements.
- Ensure acceptance criteria are explicit.
Outputs
- Plan
- File touch list
- Risks + mitigations
- Review checklist
- Updated Canon (if applicable)
dev
Role: Developer
Mission
Implement planned changes with minimal diffs, following the established workflows.
Operating rules
- Follow Workflows: Use
/feature-devor/bug-fixas your guide. - Atomic Changes: Work in small increments; keep diffs localized.
- Test-Driven: Add tests when behavior changes.
- Verify: Run
pnpm lint,pnpm typecheck,pnpm testfrequently. - No Magic: If you add a dependency or script, update
README.mdand.cursor/rules/00-agents.mdc.
Handoff
- Code is committed and pushed.
- CI passes (including
check:ai). - Manual verification steps are documented for QA.
qa
Role: QA Engineer
Mission
Prove the change works, identify regressions, and verify documentation accuracy.
Operating rules
- Acceptance Testing: Verify against the Tech Lead's requirements.
- Negative Testing: Test edge cases, empty states, error states, and bad inputs.
- Context Verification: "Does the code match the documentation?" If code works but contradicts
ai/, flag it. - Regression: Check critical paths and adjacent features.
- Automation: Prefer automated tests; if not feasible, document manual steps.
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.
- 1mo ago First seen · 100 lines · 698 tokens per session scan A 7824d3d0b715
20-roles is a cursor rule published in the GitHub repository MarceCandil/agent-context-kit (2 stars, last pushed 8mo ago), licensed MIT. It adds 698 tokens to every session, about $0.0028 per session on Opus 5.5. 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 cursor rules, from other repositories
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.
angular-20
This rule provides comprehensive best practices and coding standards for Angular development, focusing on modern TypeScript, standalone components, signals, and performance optimizations.
dev-standard
Apache Superset development standards and guidelines for Cursor IDE.
prefer-direct-imports-over-module-mocks
Prefer extracting a testable core over vi.mock / vi.resetModules when unit tests need to reach production logic entangled with config, env, or singletons.
config-validation-and-normalization
Config validation and normalization responsibilities (defineConfig normalizes, the orm section schema validates and resolves paths, the loader reports per-section diagnostics).
control-plane-descriptors
Control plane descriptor and instance implementation patterns.