README

A README describing local TorqClaw agent roles, responsibilities, model assignments, and routing rules. TorqClaw is the product-specific environment these agents are designed to operate in.

In plain words
What is it for?
Use it as a reference when routing tasks through the planner, reviewer, builder, verifier, or checklist roles in a TorqClaw workflow.
Why use it?
It records who plans, reviews, builds, verifies, and approves work, helping keep those responsibilities separate and preventing agents from using paths for another product.

Agent for Claude Code

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 agents/pilotwaffle/torqclaw/readme
Clone the repo
git clone --depth 1 https://github.com/pilotwaffle/TORQCLAW

Made for: Claude Code.

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 1,094 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 $0.01094
Opus 5 $0.00000 $0.00547
Sonnet 5 $0.00000 $0.00219
Haiku 4.5 $0.00000 $0.00109

Measured 2d ago against content hash a781d3c65c47, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

README 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.

.claude/agents/README.md · 78 lines

How it starts

The opening of the file, as written. The whole thing — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.

TorqClaw agent routing policy

Adapted from the TORQ Console harness (E:\TORQ-CONSOLE\.claude\agents). These are TorqClaw-local definitions. The Console harness drives a different product through .torq/launch/* watcher scripts and hardcoded E:\TORQ-CONSOLE\... paths; running those agents against this repo would read specs from and write verdicts into the wrong product. What is ported here is the discipline, re-pathed to E:\TorqClaw.

Roles and model assignments (operator-set, 2026-08-09)

Role Model Responsibility
G1D / planner / router Fable-5 Orchestrate. Scope the ticket, route work, hold the sequence. Does not implement.
G1R / reviewer Opus 4.7 Independent review when design, architecture, risk, or security is involved. Reviews the SPEC before the Builder spends compute.
Builder Sonnet 5 Implement the spec precisely. Does not design or approve.
G2A / final verifier Opus 4.8 Grade high-impact changes against files, tests, build output, git state, and acceptance criteria. Sole authority to authorize a push.
Cheap verification Haiku / fast Checklist grading, log triage, scope sanity checks. Never sole final authority for risky code, merges, security, production, or architecture.
Memory-writer Updates STATE.md / MEMORY.md only after meaningful progress, tests pass, and verifier/G2A passes.

Hard rules

  1. Do not stop until tests pass and verifier/G2A passes — unless blocked by missing authority, unsafe repo state, permission limits, or an operator-required decision.
  2. Never push, merge, delete, clean, reset, overwrite operator files, or run destructive commands without explicit operator approval. G2A approval is a technical precondition for a push, not operator authorization.
  3. Project-local CLAUDE.md, MEMORY.md, .claude/agents, hooks, and workflows are the active authority.

Two practices worth keeping from the Console harness

Read the full file on GitHub · 78 lines

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 · 78 lines · 0 tokens per session scan A a781d3c65c47

Subscribe to this mod's changes

README is an agent published in the GitHub repository pilotwaffle/TORQCLAW (4 stars, last pushed 8d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 1,094 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 2026-08-31.