verification-loop

A workflow for checking claims about completed work with repeatable commands and recorded results.

In plain words
What is it for?
Use it after changing runtime behavior, before releases, and when verifying MCP servers, installation wrappers, or browser-based smoke tests.
Why use it?
It reduces the risk of saying a script, skill, server, or installation works when it has not actually been tested.

Skill for Claude CodeCodex

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 skills/rexleimo/aios/verification-loop
Any agent
npx skills add rexleimo/aios --skill verification-loop
Clone the repo
git clone --depth 1 https://github.com/rexleimo/aios

Made for: Claude Code, Codex.

Per session 32 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 945 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 $0.00032 $0.00945
Opus 5 $0.00016 $0.00473
Sonnet 5 $0.00006 $0.00189
Haiku 4.5 $0.00003 $0.00094

Measured 2d ago against content hash 16cd990ddd02, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

verification-loop 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.

skill-sources/verification-loop/SKILL.md · 102 lines

How it starts

The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Verification Loop

Trigger

Use this skill when:

  • You changed runtime behavior (scripts, wrappers, MCP server, install flows)
  • You are about to say "done", "fixed", "works", or "passes"
  • You are about to bump version / release

Rules

  • Prefer commands with deterministic exit codes over "it looks fine".
  • If you cannot run verification, say exactly what you could not run and why.

Baseline Checks (AIOS)

  1. Run the verifier:

    • aios doctor
    • Or: node scripts/aios.mjs doctor
    • Compatibility wrappers: scripts/verify-aios.sh / scripts/verify-aios.ps1
  2. MCP server changes (minimum):

    • cd mcp-server && npm run typecheck
    • cd mcp-server && npm run build
    • Manual smoke: chrome.launch_cdp -> browser.connect_cdp -> page.goto -> page.extract_text/page.screenshot -> browser.close
  3. Install/wrapper changes:

    • Re-run install/update on a clean-ish shell session.
    • Confirm new commands are visible and resolve to ROOTPATH scripts.

Evidence Capture

  • Record the exact commands run and whether they succeeded.
  • For failures: include the first actionable error line and the remediation you applied.

Structured Verdict Schema

Inspired by the-pair v2.0.2 quality_gate.rs: every completion claim MUST be backed by a structured verdict with exactly four sections. A verdict missing any section is an automatic REJECT — there is no partial credit.

The four required sections, in order:

  1. FILES_REVIEWED — list of files reviewed with line ranges and change status.
  2. CHECKS — typecheck, test suite, lint — each with a concrete PASS/FAIL status.
  3. CODE — the specific code snippet or issue reference, as a quoted block.
  4. VALIDATION — summary verdict: APPROVED or REJECTED, followed by specific next_actions when rejected.

Mandatory format

VERDICT:
FILES_REVIEWED:
  - path/to/file.ts: lines 45-67 (changed)
CHECKS:
  - typecheck: PASS
  - test suite: PASS (8/8)
  - lint: PASS
CODE:
  > // specific snippet or issue reference
VALIDATION:
  APPROVED — all checks pass
  OR
  REJECTED — [specific next_actions list]

Read the full file on GitHub · 102 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 · 102 lines · 32 tokens per session scan A 16cd990ddd02

Subscribe to this mod's changes

verification-loop is a skill published in the GitHub repository rexleimo/aios (52 stars, last pushed 3d ago), licensed MIT. It adds 32 tokens to every session and 945 once invoked, about $0.0002 per session on Opus 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 2026-08-30.