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/heggria/taskflow/agents-mdgit clone --depth 1 https://github.com/heggria/taskflowWhat 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 | $0.07110 | $0.07110 |
| Opus 5 | $0.03555 | $0.03555 |
| Sonnet 5 | $0.01422 | $0.01422 |
| Haiku 4.5 | $0.00711 | $0.00711 |
Grade A, and why
taskflow 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 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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
│ │ ├─ runner.ts ← pi subagent spawn (child_process `pi --mode json`); re-exports core helpers How it starts
The opening of the file, as written. The whole thing — 318 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Instructions for AI coding agents working on taskflow.
Project Overview
taskflow is a declarative DAG orchestration runtime for coding agents — it runs on the Pi coding agent, on OpenAI Codex, on Claude Code, on OpenCode, and on Grok Build, and on Hermes Agent. It lets users define multi-phase workflows (fan-out, gate, loop, tournament, approval, sub-flow composition) as JSON DSL, executes them via isolated subagent processes, and returns only the final result — intermediate transcripts never enter the host context window.
Language: TypeScript (ES2022, ESM, --experimental-strip-types for direct execution in dev)
Runtime: Node.js ≥ 22.19 (uses fs.globSync, Atomics.wait)
Dependencies: Zero runtime deps. The private charterarc experiment depends only on taskflow-core. The Pi adapter (pi-taskflow) peer-depends on @earendil-works/pi-{agent-core,ai,coding-agent,tui}; the host-neutral MCP server (taskflow-mcp-core) and the four MCP host adapters (codex-taskflow, claude-taskflow, opencode-taskflow, grok-taskflow) all depend on taskflow-core (the adapters also depend on taskflow-mcp-core). Everything depends on typebox.
Layout: pnpm-workspace monorepo — taskflow-core (engine), private charterarc (project template loop), taskflow-mcp-core (MCP + DAG SVG), taskflow-hosts (codex/claude/opencode/grok runners), taskflow-dsl (S4: .tf.ts → Taskflow → FlowIR; CLI taskflow-dsl), pi-taskflow, codex-taskflow, claude-taskflow, opencode-taskflow, grok-taskflow, hermes-taskflow (host delivery packages).
Build: each package compiles to dist/*.js + .d.ts (tsc); published packages ship dist (Node refuses to type-strip .ts under node_modules). Dev resolves the TypeScript sources directly via a development export condition — no build needed to typecheck or test.
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 · 318 lines · 7,110 tokens per session scan A ca5827d58e76
taskflow AGENTS.md is an instructions file published in the GitHub repository heggria/taskflow (67 stars, last pushed 5d ago), licensed MIT. It adds 7,110 tokens to every session, about $0.0355 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
devspace AGENTS.md
Instructions for Waishnav/devspace, covering devspace, product model, glossary, security boundaries and diagnose the correct layer.
zenflow AGENTS.md
Instructions for zendev-sh/zenflow, covering agents.md - zenflow, commands, architecture, key rules and coordinator and messaging.
oh-my-subagents AGENTS.md
Instructions for ringlochid/oh-my-subagents, covering oh my subagents coding agent contract, product purpose, design philosophy, principles and authority split.
zenflow CLAUDE.md
Instructions for zendev-sh/zenflow, covering claude.md - zenflow, commands, architecture, key rules and coordinator and messaging.
kungfu copilot-instructions.md
Instructions for kungfu-systems/kungfu: Read and follow AGENTS.md before proposing or running changes. It is the shared agent router for this repository, including the canonical Shifu development and build entrypoint.
sofagent GEMINI.md
Instructions for KongFangXun/sofagent, covering gemini.md · sofagent gemini cli 适配（薄挂载）, 会话开始时按序加载（四层加载链）, 审计强制（平台无关） and 连接 mcp server.