Installs into .claude/skills of the current project.
Are you the author of Threat Surface Mapping?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/yogsoth-ai-threat-surface-mapping)
---
name: threat-surface-mapping
description: Enumerate all attackable surfaces of an artifact — logical, empirical, methodological, social, and practical dimensions.
execution: subagent
prompt: ./prompt.md
input: artifact (string), artifact_type (string)
used-by: [red-teaming]
---
# Threat Surface Mapping
Enumerates all surfaces of the artifact that could be attacked or challenged.
## Execution
Subagent — spawned via subagent-spawning/spawn-agent.
## Why Subagent
Surface enumeration requires systematic coverage without bias toward obvious attack points. Isolated context prevents premature focus on specific vulnerabilities.
## Input
- **artifact**: The complete artifact to analyze
- **artifact_type**: Type of artifact (hypothesis, claim, idea, approach, etc.)
## Output
- **surfaces**: List of threat surfaces with category, description, and attack accessibility rating
- **coverage_map**: Matrix of dimensions covered vs. uncovered
- **priority_ranking**: Surfaces ranked by expected vulnerability