The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from project specification files to drive the generation of design proposals and tool execution.
Ingestion points: Processes spec-product.md and tech-recon.md within the .stelow directory to determine appetite and technical constraints.
Boundary markers: Lacks explicit delimiters or instructions to the agent to ignore potentially malicious instructions embedded in the input specification files.
Capability inventory: The skill can execute shell commands (ls, grep, awk, source) and invokes external subagents and visual review tools.
Sanitization: No validation or escaping of the input file content is performed before it is used to influence the workflow logic.
[COMMAND_EXECUTION]: The skill utilizes shell commands and local script sourcing to manage its workflow and process data.
Shell utilities: Uses ls, grep, and awk to query file system state and extract text from generated artifacts.
Script sourcing: Employs the source command to load utility functions from a relative path (../../stelow-workflow-orchestrator/), establishing a dependency on an external script file from the same ecosystem.

stelow-workflow-interface-alternatives