EVIDENCE AND EVALUATION / Undominated.ai
Undominated · A missing announcement is not safe
Refuse a safety claim built from none-announced, unknown, or unrated. Those labels do not establish that a subject is safe.
“Refuse a safety claim built from none-announced, unknown, or unrated. Those labels do not establish that a subject is safe.”
01 / THE REASONING
Why this made the selection.
- Pairs a bounded workflow with an offline Python checker and explicitly synthetic example inputs.
- Created for recurring evidence and comparison failures: missing facts remain unknown and every conclusion keeps its scope.
- It names a concrete failure mode: A denial is not rewritten into a safety verdict, and a null is not a verdict either.
02 / THE REVIEW RECORD
What we actually inspected.
Source review has boundaries.
A clear record is more useful than a “safe” badge.
Material inspected
- skills/undominated-announced-not-safe/SKILL.md
- skills/undominated-announced-not-safe/scripts/check.py
- skills/undominated-announced-not-safe/examples/synthetic.json
Our findings
- First-party resource authored by Undominated.ai; this is our own verification record, not an independent endorsement.
- The download preserves the complete source and supporting files. Its manifest hashes identify the exact bytes.
- One input rule the file states is: `claimedSafe` is true, false, or null.
- One condition that blocks a pass is: label must be none-announced, unknown, or unrated.
- The checker states its scope as: none-announced, unknown, and unrated do not establish safe.
Not established by this review
- Model compliance with these instructions on arbitrary tasks
- Native integration in every agent host
- Public GitHub source identity
- Remote installation through the standard Skills CLI
- The new npm resource CLI installation
The review applies to the material and revision named here. A newer upstream release can change its behavior.
03 / PUT IT TO WORK
Add a skill to your workflow.
- Download the complete bundle from https://undominated.ai/resources/skills/undominated-announced-not-safe/bundle.zip and extract it into a new directory.
- Place the extracted directory at .agents/skills/undominated-announced-not-safe/ or your host's documented skills directory. Preserve SKILL.md, scripts, examples and licence files together.
- Run the documented synthetic example from the skill directory with Python 3.10 or later before using your own evidence.
Before you start
- An Agent Skills compatible host
- Python 3.10 or later for the optional checker
THE COMPLETE REVIEWED DEFINITION
Read it before you reuse it.
Original source bytes, with attribution.
Review the host-specific setup notes above.
---
name: undominated-announced-not-safe
description: Refuse a safety claim built from none-announced, unknown, or unrated. Those labels do not establish that a subject is safe.
license: MIT
metadata:
author: Undominated.ai
version: "1.0.0"
---
# A missing announcement is not safe
Use when a safety sentence is about to treat a missing announcement as a favourable verdict.
1. The only labels this checker reads are `none-announced`, `unknown`, and `unrated`. Any other label is invalid here.
2. `claimedSafe` true is a review. The result never sets `established` to true.
3. `claimedSafe` false or null can pass, and `established` stays false. A denial is not rewritten into a safety verdict, and a null is not a verdict either.
4. No source is fetched. The label you supply is the whole input.
## Input contract
`label` is one of the three words above. `claimedSafe` is true, false, or null. A string and the JSON number 1 are invalid.
## Deliverable and limits
Return the label, the claim you supplied, and `established: false`. A pass does not certify safety. It records that this label was not promoted to safe.
`examples/synthetic.json` passes with `none-announced` and a null claim. The true-claim example requires review.
## Run the local check
Resolve paths relative to this skill directory:
```sh
python3 scripts/check.py examples/synthetic.json
python3 scripts/check.py /absolute/path/to/your-input.json
```
The examples are **synthetic**, not a safety evaluation. Python 3.10+; standard library only. The script reads one explicit UTF-8 JSON file, prints JSON and makes no network requests or file writes.
Exit codes: `0` consistency check passed, `1` review required, `2` invalid input. CLI argument and malformed-input errors return structured JSON; `--help` displays usage text. Unknown or duplicate JSON fields, non-JSON constants and invalid types are rejected. Files are limited to 1 MiB.
This source-only skill is **not included in undominated-check@0.4.0**. Keep `SKILL.md`, the checker, examples and MIT licence together. A pass validates the bounded supplied-input contract, not production suitability or permission to publish. The skill does not authorize external actions.
The download contains SKILL.md. Extract the whole bundle; the supporting files are required. Inspect the included MANIFEST.json for file hashes.
By Undominated.ai. Exact upstream source ↗ · Licence · Attribution
SHA-256 f66189df8dc036d8d5da15ac1169908dfd3863310b514821a8340037c3273ed6
Read the applicable licence
MIT License Copyright (c) 2026 Undominated.ai Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
04 / FOLLOW THE EVIDENCE
The source trail.
Our notes are separate from the original resource.
Check upstream before adopting a new version.
https://undominated.ai/resources/skills/undominated-announced-not-safe/definition.md
Supports: Workflow and input requirements, Stated limitations
- MIT licence ↗Checked
https://undominated.ai/resources/skills/undominated-announced-not-safe/LICENSE.txt
Supports: Redistribution terms and attribution
- Complete source bundle ↗Checked
https://undominated.ai/resources/skills/undominated-announced-not-safe/bundle.zip
Supports: Full local source, supporting files and integrity manifest