The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute a suite of local management scripts (scripts/render-html.sh, scripts/plan-brief-open.sh, scripts/load-cross-project-groups.sh).\n
In Step 2 (alt), the skill instructs the agent to pass a user-provided value directly into a shell command: bash scripts/load-cross-project-groups.sh --group "<name>". This pattern creates a command injection surface if the agent does not properly sanitize the <name> argument before execution.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core functionality of ingesting and processing external data.\n
Ingestion points: Data is fetched from harness-mem search results (past decisions, patterns, and archived plans) and the user's initial task description.\n
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when processing the content of retrieved past plans.\n
Capability inventory: The skill has access to the Bash tool (script execution), Write (file creation), and Read (file access), and it can trigger a browser to open generated files.\n
Sanitization: While the skill mentions a redaction layer (--with-redaction) to prevent PII leakage during cross-project searches, it does not describe sanitization or filtering logic specifically designed to prevent malicious instructions embedded in project memory from influencing agent behavior.

harness-plan-brief