The Agent Skills Directory

[PROMPT_INJECTION]: No override or bypass markers detected in the skill instructions or metadata.
[DATA_EXPOSURE_AND_EXFILTRATION]: No access to sensitive file paths or unauthorized network operations were identified.
[OBFUSCATION]: No Base64, hex-encoded, or hidden content detected in the scripts or documentation.
[UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: No external script downloads or package installations found in the provided code.
[PRIVILEGE_ESCALATION]: The scripts/init_skill.py script sets executable permissions (0o755) on generated template scripts, which is standard behavior for project scaffolding tools.
[PERSISTENCE_MECHANISMS]: No attempts to establish persistent access via shell profiles, cron jobs, or startup tasks.
[METADATA_POISONING]: Skill metadata is descriptive and aligned with the stated functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided strings through CLI arguments to generate files. 1. Ingestion points: CLI arguments in init_skill.py and generate_openai_yaml.py. 2. Boundary markers: None. 3. Capability inventory: File system writes, directory creation, and permission changes. 4. Sanitization: Normalization of skill names and escaping of YAML values to prevent format injection.
[TIME_DELAYED_OR_CONDITIONAL_ATTACKS]: No logic gating behavior based on time or environment detected.
[DYNAMIC_EXECUTION]: No runtime code generation or unsafe deserialization of external data was found.

skill-creator