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/levi-qiao/longgraph-skill/agents-mdgit clone --depth 1 https://github.com/levi-qiao/longgraph-skillWrote 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/levi-qiao/longgraph-skill/agents-md)<a href="https://agentmods.dev/instructions/levi-qiao/longgraph-skill/agents-md"><img src="https://agentmods.dev/badge/instructions/levi-qiao/longgraph-skill/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 | $0.01563 | $0.01563 |
| Opus 5 | $0.00781 | $0.00781 |
| Sonnet 5 | $0.00313 | $0.00313 |
| Haiku 4.5 | $0.00156 | $0.00156 |
Grade A, and why
longgraph-skill 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 4d 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.
- **Distribution**: installable via the curl script *and* the Claude Code plugin. The How it starts
The opening of the file, as written. The whole thing — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — working on longgraph-skill
Guidance for any agent (or human) editing this repo. Read before changing anything.
The same rules apply in Claude Code — CLAUDE.md imports this file.
What this repo is
A curated, opinionated prompt library for long-horizon agent work, framed as
"graph engineering." It ships Markdown prompts, not application code — there
is no build step and no runtime. One umbrella (/longgraph) checks whether the
method fits, then delegates to the loop-graph compiler. Specialized entries such as
skills/loop-converge/, skills/loop-deliver/, and
skills/loop-research/ bind a goal pack and follow that same compiler — they do
not add a runtime.
skills/loop-graph/— an executor node + a separate-context supervisor node, each on its own recurring timer, with no wake edge between them.skills/loop-converge/— preset entry: code-convergence interview → same compile.skills/loop-deliver/— preset entry: requirement-delivery interview → same compile.skills/loop-research/— preset entry: evidence-led solution selection → same compile.
Simple self-contained goals use the host's ordinary task/goal directly; longgraph does not wrap them in a second objective.
Deep context lives in lib/methodology.md (the why behind each
rule) and skills/loop-graph/docs/model.md (the
node/edge vocabulary + invariants). The contribution bar is in CONTRIBUTING.md.
Layout
| Path | What it is |
|---|---|
SKILL.md |
/longgraph router and authoring entrypoint; may launch nodes but never acts as one |
skills/loop-graph/SKILL.md |
shared compiler: interview → generate → deliver |
skills/loop-{converge,deliver,research}/ |
focused preset entries: bind their goal pack → same compile |
skills/loop-graph/templates/*.md |
compiled runtime prompts — not code; keep {{PLACEHOLDER}}s and structural headings intact |
skills/loop-graph/docs/preset-contract.md |
one source of truth for baseline versus goal-pack ownership |
skills/loop-graph/examples/ |
concrete, fully worked runs (these are project-specific — that's correct) |
skills/loop-graph/references/*.md |
one independently loaded owner per host (invocation, context, hooks, handoff fields) |
.claude-plugin/ |
Claude Code plugin + marketplace manifests |
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.
- 4d ago First seen · 99 lines · 1,563 tokens per session scan A 71e89f499ac6
longgraph-skill AGENTS.md is an instructions file published in the GitHub repository levi-qiao/longgraph-skill (69 stars, last pushed yesterday), licensed MIT. It adds 1,563 tokens to every session, about $0.0078 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
paperclip-plugin-company-wizard CLAUDE.md
Instructions for Yesterday-AI/paperclip-plugin-company-wizard, covering claude.md, what this is, commands, architecture and source layout.
paperclip-plugin-company-wizard AGENTS.md
Instructions for Yesterday-AI/paperclip-plugin-company-wizard, covering company wizard — project vision, what, why, core idea: gracefully optimistic architecture and design principles.
team-shinchan AGENTS.md
Instructions for seokan-jeong/team-shinchan, covering agents.md - team-shinchan agent map, layer architecture, agent registry, call flow and quick reference.
team-shinchan CLAUDE.md
Instructions for seokan-jeong/team-shinchan, covering team-shinchan plugin, repository-specific notes and source layer (src/).
claude-mpm CLAUDE.md
Claude Code instructions for bobmatnyc/claude-mpm, covering claude mpm project, priority guide, 🔴 git commit rules, 🟡 worktree development workflow and 🔴 release workflow.
claude-code-plugin CLAUDE.md
Instructions for anilcancakir/claude-code-plugin, covering claude.md, mission, architecture, commands (ac plugin) and skills & mcp.