stream-chain
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements a sequential pipeline where the output of each agent is fed directly into the prompt context of the next agent. This architecture allows for potential indirect prompt injection if an intermediate step generates or processes untrusted content that contains adversarial instructions.\n
- Ingestion points: Sequential output flow from steps in
runandpipelinemodes as described in SKILL.md.\n - Boundary markers: No explicit delimiters or boundary markers (e.g., XML tags or specific 'ignore instructions' prefixes) are defined in the orchestration logic or context flow description.\n
- Capability inventory: Code analysis, automated refactoring, test generation, and performance optimization via the
claude-flowCLI as specified in the pipeline definitions.\n - Sanitization: The skill does not appear to implement sanitization or validation of the text output between steps in a chain.\n- [COMMAND_EXECUTION]: The skill defines and uses command-line patterns for the
claude-flowutility to manage agents and workflows. These commands are integral to the skill's functionality as a development tool and follow the platform's standard usage patterns.
Audit Metadata