mythic-translation-containers
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions focus on legitimate development tasks for the Mythic C2 framework. It includes specific security recommendations such as constant-time comparisons, failing closed on unknown crypto values, and avoiding the logging of sensitive data.\n- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow for processing untrusted agent wire messages. While it provides security checklists to mitigate risks, the resulting translation container by definition handles external, potentially attacker-controlled input.\n
- Ingestion points: Agent wire messages (custom format) processed by translation functions as described in SKILL.md and references/translation-workflow.md.\n
- Boundary markers: The skill instructs to 'Reject malformed messages before parsing deeply' and 'Verify MAC/signature' in references/translation-workflow.md.\n
- Capability inventory: Translation containers interact with the Mythic C2 ecosystem for message conversion and key generation.\n
- Sanitization: Instructions include validating crypto parameters and failing closed on unknown values in SKILL.md.
Audit Metadata