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 jaktestowac/awesome-copilot-for-testers --skill reviewing-ai-output-groundednessgit clone --depth 1 https://github.com/jaktestowac/awesome-copilot-for-testersWrote 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/jaktestowac/awesome-copilot-for-testers/reviewing-ai-output-groundedness)<a href="https://agentmods.dev/skills/jaktestowac/awesome-copilot-for-testers/reviewing-ai-output-groundedness"><img src="https://agentmods.dev/badge/skills/jaktestowac/awesome-copilot-for-testers/reviewing-ai-output-groundedness/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/jaktestowac/awesome-copilot-for-testers/reviewing-ai-output-groundedness"><img src="https://agentmods.dev/badge/skills/jaktestowac/awesome-copilot-for-testers/reviewing-ai-output-groundedness.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 | $0.00103 | $0.02179 |
| Opus 5 | $0.00051 | $0.01090 |
| Sonnet 5 | $0.00021 | $0.00436 |
| Haiku 4.5 | $0.00010 | $0.00218 |
Grade A, and why
reviewing-ai-output-groundedness 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 12d 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 — 152 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Reviewing AI Output Groundedness
Use this skill when a feature states facts derived from sources, and someone has to establish whether those facts are actually in the sources.
An eval suite catches structural regressions and can approximate faithfulness with a judge. It cannot tell you whether a confident, fluent, well-cited paragraph is quietly wrong in a way that matters to this domain. That judgement is human, it is sampled rather than exhaustive, and it ends in an attestation rather than a pass.
Fluency is the trap. Ungrounded output does not look broken - it looks better than grounded output, because nothing in it is hedged.
When to Use
- a RAG, summarisation, extraction, or research feature is going in front of users
- output is used to make a decision: medical, legal, financial, operational
- an eval suite reports high faithfulness and someone is not convinced
- users report answers that were "wrong but sounded right"
- a release needs a factuality attestation, not just a score
- a responsible-AI review is required before launch
Operating Principles
- A claim is the unit, not the answer. A four-sentence answer contains eight claims, seven grounded and one invented. Reviewing "the answer" misses it every time.
- Grounded means present in the retrieved source. Not "true", not "plausible", not "consistent with what I know". A claim that is true but absent from the sources is still ungrounded - the system got lucky.
- Citations are checked, not counted. A citation that does not support the claim it is attached to is worse than none: it manufactures confidence.
- Sample defensibly and say how. A convenience sample of ten happy-path outputs supports no conclusion. State the size, the selection method, and the confidence you are claiming.
- Classify, do not just flag. Fabrication, unsupported inference, conflation, stale source, and omission need different fixes; a single "hallucination" count directs nobody.
- Omission is a groundedness failure. A summary that drops the one caveat that mattered is faithful to the source and useless to the reader.
- Absence of a source is a required behaviour. "The documents do not say" is a correct answer, and a feature that cannot produce it will invent instead.
- The output is an attestation with limitations. Never a percentage presented as a guarantee.
What ships with it
2 files 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.
- 12d ago First seen · 152 lines · 103 tokens per session scan A 23ea97cf09a9
reviewing-ai-output-groundedness is a skill published in the GitHub repository jaktestowac/awesome-copilot-for-testers (113 stars, last pushed 16d ago), licensed MIT. It adds 103 tokens to every session and 2,179 once invoked, about $0.0005 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.
Other skills, from other repositories
standalone-python-scripts
Skill "standalone-python-scripts" from iloveitaly/llm-ide-rules, covering standalone python scripts, /// script, requires-python = ">=3.13", dependencies = [] and ///.
skill-meta-prompt
Craft better prompts using proven optimization techniques — use when your prompt needs refinement.
dashscope
DashScope (Alibaba Cloud Bailian / 阿里云百炼) integration — image generation (qwen-image-2.0-pro), text-to-speech (qwen3-tts-flash), and ASR with word-level timestamps (qwen3-asr-flash-filetrans). Use when generating images via Qwen-Image, narrating via Qwen-TTS, or transcribing with word-level timestamps via Qwen-ASR.
microsoft-extensions-ai
Build provider-agnostic .NET AI integrations with Microsoft.Extensions.AI, IChatClient, embeddings, middleware, structured output, vector search, and evaluation. USE FOR: building or reviewing .NET code that uses Microsoft.Extensions.AI, Microsoft.Extensions.AI.Abstractions, IChatClient, IEmbeddingGenerator…
sep
Use Sep for high-performance separated-value parsing and writing in .NET, including delimiter inference, explicit parser/writer options, and low-allocation row/column workflows. USE FOR: delimited data needs are performance-sensitive and allocation-aware; project needs explicit control over separator inference…
managedcode-markitdown
Use ManagedCode.MarkItDown when a .NET application needs deterministic document-to-Markdown conversion for ingestion, indexing, summarization, or content-processing workflows. USE FOR: ManagedCode.MarkItDown integration; document ingestion flows; Office or rich-text conversion to Markdown; indexing and summarization…