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.
npx skills add OutlineDriven/odin-claude-plugin --skill property-test-failure-triagegit clone --depth 1 https://github.com/OutlineDriven/odin-claude-pluginWrote 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/skills/outlinedriven/odin-claude-plugin/property-test-failure-triage)<a href="https://agentmods.dev/skills/outlinedriven/odin-claude-plugin/property-test-failure-triage"><img src="https://agentmods.dev/badge/skills/outlinedriven/odin-claude-plugin/property-test-failure-triage/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/skills/outlinedriven/odin-claude-plugin/property-test-failure-triage"><img src="https://agentmods.dev/badge/skills/outlinedriven/odin-claude-plugin/property-test-failure-triage.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00056 | $0.01265 |
| Opus 5 | $0.00028 | $0.00633 |
| Sonnet 5 | $0.00011 | $0.00253 |
| Haiku 4.5 | $0.00006 | $0.00127 |
Grade A, and why
property-test-failure-triage 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 7d 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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Property test failure triage
Contract
| Field | Bound contract |
|---|---|
| Trigger | A generated or shrunk counterexample exists and the user asks whether it reveals a code bug, an over-broad strategy, an incorrect property, or an unsettled specification. |
| Authority | Read-only: no file, VCS, credential, paid, published, deployed, or remote mutation. |
| Side effect | In chat, investigate the failing input and contract, classify the failure, and recommend the minimal code, strategy, property, or specification action. |
| Done | The counterexample is classified with cited evidence and a concrete next action; any ambiguity is left for the maintainer rather than mislabeled as a defect. |
Inputs
- Required: the failing counterexample value (original or shrunk). If the user supplies a shrunk counterexample, also supply the original generated value when available.
- Required: the property-based test that produced the failure, accessible for reading.
- Required: the implementation code the property tests.
- Optional: the test framework and shrinker name.
- Optional: the specification or requirements document, when one exists.
Procedure
- Confirm the failing counterexample is non-empty and represents a concrete execution trace. If the counterexample is absent or is only a framework-generated failure with no concrete input, return
failure-input-missing. Done when: the counterexample is confirmed as a concrete execution trace orfailure-input-missingis returned. - Identify the property-based test framework and shrinker in use from the test file or build configuration. If the test file cannot be read, return
test-file-unreadable. Done when: the framework and shrinker are identified ortest-file-unreadableis returned. - Read the complete property function and its surrounding test scaffold. Extract:
- the property statement (the predicate that evaluated to false);
- the generators or data sources used to produce the counterexample;
- any custom shrink configuration.
If the property cannot be extracted, return
property-unreadable. Done when: the property statement, generators, and shrink configuration are extracted orproperty-unreadableis returned.
- Read the implementation code under test. Build a minimal reproduction of the failure by substituting the counterexample into the code path the property exercises. If the implementation cannot be read, return
implementation-unreadable. Done when: a minimal reproduction is built orimplementation-unreadableis returned. - Classify the failure using the following mutually exclusive categories, in precedence order: a. Code Bug — the implementation produces an incorrect result or side effect for the counterexample, and the property statement correctly describes the intended behavior. The minimal repair is a code change to the implementation. b. Over-broad Property — the property predicate rejects a value that the implementation is permitted to produce under the current specification. The minimal repair is to narrow the property to the set of values the implementation actually guarantees. c. Incorrect Property — the property predicate describes behavior the implementation does not claim to guarantee, or the predicate is logically wrong independent of the implementation. The minimal repair is to correct or remove the property. d. Unsettled Specification — neither the implementation nor the property can be declared wrong because the requirement itself is ambiguous, absent, or contested. The minimal repair is to resolve the specification with the maintainer before touching code or property. Done when: the failure is classified into exactly one category with cited evidence.
- Cite the specific evidence for each category:
- For Code Bug: quote the exact implementation behavior for the counterexample, quote the property predicate, and state which implementation branch or operation produces the incorrect result.
- For Over-broad Property: quote the relevant specification clause or implementation comment that permits the counterexample outcome.
- For Incorrect Property: quote the property predicate and demonstrate the logical inconsistency with the implementation's documented contract.
- For Unsettled Specification: name the absent or ambiguous requirement and state what must be decided before any code or property change is valid. Done when: the specific evidence for the chosen category is cited.
- State the minimal repair action in one concrete imperative sentence. Do not add extra refactorings, style changes, or tests beyond the minimum required to make the property pass or correctly fail. Done when: the minimal repair is stated as one concrete imperative sentence.
- If evidence supports more than one category, report the ambiguity explicitly with the competing classifications and let the maintainer decide. Do not pick one classification to force a resolution. Done when: ambiguity is reported with competing classifications or no ambiguity exists.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 7d ago First seen · 66 lines · 56 tokens per session scan A f6c3d6954a1c
property-test-failure-triage is a skill published in the GitHub repository OutlineDriven/odin-claude-plugin (35 stars, last pushed 2d ago), licensed Apache-2.0. It adds 56 tokens to every session and 1,265 once invoked, about $0.0003 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 2026-09-04.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…