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 rules/gwaghmar/graph/graphgit clone --depth 1 https://github.com/gwaghmar/graphWhat 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 | $0.00015 | $0.00296 |
| Opus 5 | $0.00008 | $0.00148 |
| Sonnet 5 | $0.00003 | $0.00059 |
| Haiku 4.5 | $0.00002 | $0.00030 |
Grade A, and why
graph 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.
What it actually says
Graph workflow
When the user invokes @graph:
- Size first: estimate files to change (F), design decisions (D), cross-cutting risk (R). F ≤ 1 with no D/R → one implementation node plus local quality checks, no planner or reviewer agents. F 2–4 or D = 1 → 2–3 nodes, at most one reviewer. F ≥ 5, D ≥ 2, or R → full planner/workers/reviewer protocol, one worker per independent file cluster.
- Initialize a run, or resume the latest incomplete run when requested.
- Record planning, implementation, checks, review, and synthesis nodes, then run
.graph/graph.py validateand fix the plan if it reports unknown or cyclic dependencies. - Check the local content cache before any avoidable model call.
- Run local quality gates before model reviewers.
- Retry only failed nodes and dependent nodes, maximum twice.
- Cache successful reusable node outputs.
- Rely on the live ASCII graph printed by each
.graph/graph.py nodecall as the progress display; do not re-describe it in prose. - Keep the local HTML graph updated; at completion, run
.graph/graph.py finishand relay the printed run summary plus the report path. - Commit only with explicit user permission.
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 · 18 lines · 15 tokens per session scan A 8463a25d4a70
graph is a cursor rule published in the GitHub repository gwaghmar/graph (10 stars, last pushed 23d ago), licensed MIT. It adds 15 tokens to every session and 296 once invoked, about $0.0001 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 cursor rules, from other repositories
quiet-harness
QuietHarness baseline.
harness-workflow-overview
Harness Workflow overview for Cursor project rules.
harness-builder
Build or repair a project-level AI-agent workbench; route thick sub-capabilities to helper skills.
plan
Use plan when an approved Spec or clear non-trivial request must become an Executable Plan.
brainstorm
Use brainstorm when requirements, boundaries, solution tradeoffs, success criteria, or verification strategy are not yet settled.
diagnose
Use diagnose when build, test, lint, typecheck, CI, runtime, or smoke checks fail without a proven root cause.