validation-review

A read-only review of how easily a small code change can be verified. It examines the repository’s available checks and runs the most relevant path rather than assuming the entire project must be tested.

In plain words
What is it for?
Use it before relying on a test or build gate, especially when checking JavaScript, TypeScript, Python tests, or API payload changes.
Why use it?
It reveals whether validation is targeted and trustworthy, or slow, noisy, incomplete, or passing without checking the changed behavior.

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/stroland02/sync/validation-review
Clone the repo
git clone --depth 1 https://github.com/stroland02/Sync

Made for: Claude Code.

Per session 41 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 665 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 $0.00041 $0.00665
Opus 5 $0.00020 $0.00332
Sonnet 5 $0.00008 $0.00133
Haiku 4.5 $0.00004 $0.00067

Measured yesterday against content hash 36ea114bba47, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

validation-review 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 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.

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/validation-review.md · 56 lines

How it starts

The opening of the file, as written. The whole thing — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Validation review

Checks whether an agent can verify a small change without falling back to a full-repository loop.

Read-only. Run the checks; never edit a file to make one pass.

Workflow

  1. Inspect the declared paths: web/package.json scripts, the pytest suites under tests/, oxlint, npx tsc -b, and the gate CLAUDE.md names as the authority.
  2. Decide whether a scoped loop exists for a small change, or whether every change pays for the whole tree.
  3. Run the most relevant path and time it.
  4. Judge the result: targeted, actionable, noisy, or too expensive for normal iteration.

What this repository already knows, and what to check against it

  • CI is not the authority here. Hosted runners report a job that never started as failure (B112), so the local gate is what counts. A validation loop that only exists in CI scores badly.
  • A green JavaScript gate is not evidence. npm run build typechecks the console against the types the console declares, not against what the API sends. Judge whether the loop can catch a payload change, not merely a syntax error.
  • Watch for a loop that passes vacuously. Guards here have twice been green for the wrong reason: an assertion that a band is absent passed because a prior test's DOM was still mounted, and a queryByLabelText passed because the component early-returns on empty data whether or not the page gates it. A loop that cannot fail is worth less than no loop.
  • Report honestly if a suite is slow: the full Python suite runs into the tens of minutes serially, and whether a scoped subset exists is exactly what this review is for.

Scoring

  • around 93/100 if there is a repeatable path giving useful signal, even if broader than ideal.
  • around 84/100 if validation works but is heavier than it should be, or split across commands.
  • around 68/100 if a loop probably exists but choosing it takes guesswork, or the output is too noisy to trust quickly.
  • around 27/100 if there is no practical loop you can actually use.
  • around 12/100 if it is blocked on secrets or infrastructure you cannot reach.

Read the full file on GitHub · 56 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. yesterday First seen · 56 lines · 41 tokens per session scan A 36ea114bba47

Subscribe to this mod's changes

validation-review is an agent published in the GitHub repository stroland02/Sync (5 stars, last pushed 5d ago), licensed Apache-2.0. It adds 41 tokens to every session and 665 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 2026-08-31.