The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using python3 to run external scripts defined by environment variables ($CATC_MCP_SCRIPT, $MCP_CALL, $PYATS_MCP_SCRIPT). Shell commands are constructed by concatenating environment variables and JSON-formatted tool arguments, which could be vulnerable to command injection if input is not strictly validated.
[CREDENTIALS_UNSAFE]: Sensitive credentials (CCC_PWD) are passed directly as environment variables within shell command invocations (e.g., CCC_PWD=$CCC_PWD python3 ...). This can lead to credential exposure in system process logs or process monitors.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from network devices via Catalyst Center APIs and pyATS show commands. 1. Ingestion points: fetch_devices, fetch_interfaces, and pyats_run_show_command (SKILL.md). 2. Boundary markers: The instructions lack explicit delimiters or warnings to ignore embedded instructions in the fetched data. 3. Capability inventory: The agent has the capability to execute shell commands via python3 and record audit logs via the GAIT tool (SKILL.md). 4. Sanitization: There are no instructions provided to sanitize or escape device-provided data before it is processed or reported.

catc-inventory