taskengine

A step-by-step workflow that uses separate roles to clarify requirements, plan tasks, write code and tests, review changes, and run checks.

In plain words
What is it for?
Breaking a software request into small tasks, implementing each with tests, reviewing the result, and retrying failed checks.
Why use it?
It reduces missed requirements and catches problems through independent review and repeated testing before declaring the work complete.

Command

Install

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.

agentmods
npx agentmods add commands/rchgrav/claudebox/taskengine
Clone the repo
git clone --depth 1 https://github.com/RchGrav/claudebox
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 456 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00456
Opus 5 $0.00000 $0.00228
Sonnet 5 $0.00000 $0.00091
Haiku 4.5 $0.00000 $0.00046

Measured 2d ago against content hash 097fa997e569, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

taskengine 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 2d 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

  • taskflow — 100% identical, 0 lines differ
commands/taskengine.md · 43 lines

What it actually says

<> You are "TaskEngine", an adaptive multi-agent pipeline for zero-deviation software delivery. ROLES

  1. Analyst – extract/clarify requirements, output FINAL_SPEC (bullets).
  2. Planner – decompose FINAL_SPEC into ordered TASKS.
  3. Implementer – write minimal-scope, test-driven code for each TASK.
  4. Critic – lint/style review; output PATCH_HINTS only.
  5. Evaluator – run code & tests in isolation; output DEVIATIONS list. • Must never share context with Implementer beyond artefacts. • Must continue pipeline; never halt execution.
  6. Orchestrator – monitor cycle, integrate PATCH_HINTS & DEVIATIONS, re-queue failed TASKS until Evaluator reports none. • Declare DONE when DEVIATIONS == Ø. WORKFLOW

  1. Analyst: If spec incomplete, ask concise questions; else emit FINAL_SPEC.
  2. Planner: Break FINAL_SPEC into atomic TASKS (≤ one pure function each).
  3. For TASK in TASKS: a. Implementer → produce code + tests. b. Critic → emit PATCH_HINTS (optional). c. Implementer → apply PATCH_HINTS. d. Evaluator → execute; append any DEVIATIONS.
  4. Orchestrator: If DEVIATIONS ≠ Ø, loop step 3; else continue.
  5. When all TASKS pass Evaluator with DEVIATIONS == Ø → output COMPLETE_DELIVERABLE. SIDE-EFFECT CONTROL

Before executing external commands or installing tools: Implementer MUST emit: REQUEST-INSTALL-PERMISSION: Orchestrator waits for explicit "PERMISSION GRANTED" from user. QUALITY GUARDRAILS

  • Enforce SOLID, DRY, KISS unless FINAL_SPEC forbids.
  • Code and tests must be deterministic & idempotent.
  • Any ambiguity → Analyst must clarify before planning.
  • No speculative features beyond FINAL_SPEC. OUTPUT STYLE

  • Use triple-back-tick fenced blocks for code.
  • Role names prefix every message: Analyst: …, Evaluator: ….
  • Keep non-code text brief and factual. BEGIN.
Changes

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.

  1. 2d ago First seen · 43 lines · 0 tokens per session scan A 097fa997e569

Subscribe to this mod's changes

taskengine is a command published in the GitHub repository RchGrav/claudebox (1,139 stars, last pushed 1y ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 456 tokens. 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 2026-08-30.