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.
git clone --depth 1 https://github.com/MadaraUchiha-314/the-loopWrote 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/commands/madarauchiha-314/the-loop/verify-work)<a href="https://agentmods.dev/commands/madarauchiha-314/the-loop/verify-work"><img src="https://agentmods.dev/badge/commands/madarauchiha-314/the-loop/verify-work/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/madarauchiha-314/the-loop/verify-work"><img src="https://agentmods.dev/badge/commands/madarauchiha-314/the-loop/verify-work.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00025 | $0.00981 |
| Opus 5.5 | $0.00010 | $0.00392 |
| Sonnet 5.5 | $0.00005 | $0.00196 |
| Haiku 4.5 | $0.00003 | $0.00098 |
Grade A, and why
verify-work 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
the-loop: verify-work $ARGUMENTS
Prove the work item: execute the locked testing-plan.md and turn it into a record —
the verification node, between implementation and the review chain. A slice of
/the-loop:work-on (step 3 of /the-loop:execute-tasks); work-on remains the superset.
Read the the-loop skill and reference/testing.md first. Load
.the-loop/harness-config.yaml and every doc registered in customInstructions.docs —
the environment the plan references is usually described there.
Steps
-
Locate & load. Resolve
$ARGUMENTStodocs/specs/<id>/and readtesting-plan.mdandtasks.md. Every task should be ticked; if the DAG is unfinished, say so and stop — verification proves finished work.No
testing-plan.md? Thentest-planningwas declared skipped for this work item (issue-179) — checkwork-item-state.json. Do not author one now: the phase was declared away by a human, and re-creating its artifact from this side of the gate is exactly what that declaration refused. Verify the change on its merits instead (run the project's linters/tests, exercise what changed), and record the results inevidence/verification.md's## Verification resultssection — the node gates that file in place of the plan and blocks until it holds something. Everything below about what to run and how to record it still applies; only the file changes. -
Bring up the environment declared in the plan's Verification environment: repositories, services, fixtures, using the project's own commands. Credentials resolve from the referenced env vars / secret store — the plan never contains values. If bring-up fails: record it under Verification results, leave the dependent activities unticked, and escalate. Do not pass the gate on an environment that never came up.
-
Run each planned activity in the plan's Verification activities checklist. Tick a line (
- [ ]→- [x]) only once it has actually run and its evidence is recorded. An activity that cannot be executed stays unticked: write why under Verification results, then either replan the matrix (with the reason) or escalate. Silently dropping a planned activity is not permitted.
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.
- 1mo ago Changed 10d4d736bb0a
- 1mo ago Changed a12bf6fa8a17
- 1mo ago First seen · 67 lines · 25 tokens per session scan A 1274f0fa64e2
verify-work is a command published in the GitHub repository MadaraUchiha-314/the-loop (7 stars, last pushed 4d ago), licensed MIT. It adds 25 tokens to every session and 981 once invoked, about $0.0001 per session on Opus 5.5. 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 commands, from other repositories
tdd
Guide a Test-Driven Development workflow for the current task.
agent-eval-design
Design an evaluation for an AI agent or LLM feature: what to test, how to grade it, and how to catch regressions.
plan-refactor
Plan a refactor as a sequence of safe verified steps, with tests established before anything moves.
prompt-optimization
Improve an underperforming prompt through structured changes measured against fixed cases.
ijfw-audit
Run the IJFW audit gate for the current workflow phase. Usage: /ijfw-audit [phase name].
api-aqa-flow-execution-and-report-analysis
Phase 6 Execution & Report Analysis of api-aqa-flow (USER INTERACTION REQUIRED).