The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/post-threads.sh executes a Python script via a hardcoded absolute path: /Users/tomas/.openclaw/workspace/scripts/utils/clean_md.py. This references a file outside the skill's directory and assumes a specific file system structure, which can lead to the execution of untrusted or malicious code if the path is compromised or if the script is run in an environment where that path is attacker-controlled.
[DATA_EXFILTRATION]: The skill is designed to upload local images and videos to Imgur (api.imgur.com) to generate the public URLs required by the Meta Graph API. It also handles sensitive environment variables, including INSTAGRAM_ACCESS_TOKEN and THREADS_ACCESS_TOKEN. These credentials and media files are sent to external services (Meta and Imgur), which is consistent with the skill's stated purpose but constitutes a significant data flow to third-party endpoints.
[PROMPT_INJECTION]: The skill accepts user-provided files as input for captions and media. There are no clear boundary markers or sanitization steps implemented in the shell scripts to isolate this untrusted content, which creates a surface for indirect prompt injection if the resulting social media posts are subsequently processed by other AI agents.

Instagram API