explainability skill
Perform model assessment tasks including saliency sanity checks, localisation, and attribution framing.
- GitHub stars
- 378
- Files
- 1
- Last commit
- Nov 10, 2026
- License
- MIT
- Allowed tools
- —
What does the explainability skill do?
explainability is an agent skill. Perform model assessment tasks including saliency sanity checks, localisation, and attribution framing.
Summary generated from the skill's SKILL.md.
About the explainability skill
Performing model assessment tasks like saliency sanity checks, localisation, and attribution framing is what this skill enables for developers. It provides a structured way to evaluate how models make decisions by focusing on specific phases of assessment.
An agent should load this skill when a user requires deep inspection of model behavior through technical validation methods. The tool relies on the specific workflows found in Part D, Phases 12 and 13 of the model assessment process. Users should note that while the command was previously known as /explainability, it has been renamed to /model-assessment in version 6.0 and the redirect is removed in version 7.0.
What an agent can do with it
- 1Conduct saliency sanity checksA developer can ask an agent to verify the reliability of saliency maps used in model interpretation.
- 2Perform model localisationAn AI client can be tasked with identifying specific areas of interest within a model's decision process.
- 3Execute attribution framingUsers can request the framing of attributions to better understand model outputs.
Written from the skill's SKILL.md. Every use case is checked against a quote from it.
How to install the explainability skill
Each command copies the skill folder into the client's skills directory. Restart the client or start a new session to load it.Install the explainability skill in Plugin
/plugin marketplace add Aperivue/medsci-skills /plugin install medsci-modeling@medsci-skills
Install the explainability skill in Claude Code
~/.claude/skills/explainability/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/explainability ~/.claude/skills/explainabilityInstall the explainability skill in Claude Code (project)
.claude/skills/explainability/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/explainability .claude/skills/explainabilityInstall the explainability skill in Codex
~/.agents/skills/explainability/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/explainability ~/.agents/skills/explainabilityInstall the explainability skill in Gemini CLI
~/.gemini/skills/explainability/SKILL.md
$ npx degit Aperivue/medsci-skills/skills/explainability ~/.gemini/skills/explainabilityInstall the explainability skill in Claude.ai
npx degit Aperivue/medsci-skills/skills/explainability explainability && zip -r explainability.zip explainability # claude.ai → Customize → Skills → + → Create skill → Upload a skill → explainability.zip
SKILL.md
View on GitHub/explainability was renamed to /model-assessment in v6.0; this redirect is removed in v7.
Run /model-assessment now, passing the same arguments the user gave /explainability.
Start at /model-assessment Part D, Phases 12–13 (saliency sanity checks, localisation, attribution framing).
Files in this skill
1explainability skill questions
How do I install the explainability skill in Claude Code?
Run npx degit Aperivue/medsci-skills/skills/explainability ~/.claude/skills/explainability. Claude Code loads skills from ~/.claude/skills (all projects) or .claude/skills (one project) and uses a skill when its description matches the task.
Which clients can use the explainability skill?
Any client that reads the Agent Skills format: Claude Code, Codex, Gemini CLI and claude.ai (upload as a ZIP). Scripts in the skill run only where the client allows code execution.
What happened to the /explainability command?
The /explainability command was renamed to /model-assessment in version 6.0, and the redirect was removed in version 7.
How do I run model assessment tasks?
You should run the /model-assessment command using the same arguments previously used for /explainability, starting at Part D, Phases 12–13.