The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interacts with the Meego platform by executing shell commands through the meegle CLI as documented in SKILL.md and references/cli-guide.md. This is the primary and intended mechanism for the skill's functionality.
[DATA_EXPOSURE]: The skill handles project data and attachments. It includes specific commands for uploading and downloading files (meegle attachment +upload, meegle attachment +download), which are restricted to the context of Meego work items.
[PROMPT_INJECTION]: An indirect prompt injection surface is present because the skill processes content from work item descriptions and comments that could contain instructions for the agent.
Ingestion points: Data enters the context through workitem get, workitem query, and comment list commands as described in SKILL.md and references/misc.md.
Boundary markers: The skill employs structured MQL syntax and strict field-value stringification protocols which serve as logical boundaries between instructions and data.
Capability inventory: The agent has capabilities to update items, transition nodes, and upload files using the meegle CLI.
Sanitization: While the SOPs include rigorous validation of field formats and error self-healing, they do not explicitly specify sanitization of natural language content for embedded instructions.
[SAFE]: The skill is authored by larksuite, a recognized vendor. All referenced resources, including domains like meegle.com and feishu.cn, are consistent with the vendor's official infrastructure. No malicious patterns or unauthorized exfiltration attempts were detected.

meegle