sequential-thinking
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of instructional markdown files defining reasoning methodologies and deep-thinking guidelines. No executable code, malicious logic, or prompt injection attempts were found in the provided files.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface by instructing the agent to ingest and analyze external data such as logs, metrics, and file contents. However, the skill provides robust mitigation strategies to ensure data is verified and reasoning is grounded.
- Ingestion points: Diagnostic data analysis mentioned in SKILL.md and references/root-cause-playbook.md.
- Boundary markers: The skill encourages explicit knowledge classification into Verified, Believed, and Speculative categories to create logical boundaries for information trust, as detailed in references/grounding.md.
- Capability inventory: The skill utilizes Read, Write, Bash (*), and the mcp__sequential-thinking__sequentialthinking tool as defined in the SKILL.md frontmatter.
- Sanitization: The framework mandates logical falsifiability testing and cross-verification using multi-approach synthesis to validate interpretations of external data.
Audit Metadata