soia-cwork-feishu-doc-git-sync
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
subprocess.runinscripts/sync_feishu_wiki.pyto invokelark-cli. This is the primary method used to communicate with the Feishu API and is consistent with the skill's stated purpose. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted content from external Feishu documents.
- Ingestion points: Content is retrieved from Feishu via the
lark-cliutility as documented inreferences/block-mapping.yml. - Boundary markers: The skill employs HTML comments and specific labels (e.g., '嵌入式表格') to delimit mirrored content, as described in
references/sheet-mirroring.yml. - Capability inventory: The skill has the ability to write to the local file system to create the Markdown mirror and execute shell commands via
subprocess.run. - Sanitization: The skill performs normalization of Feishu-specific XML/HTML tags into standard Markdown and redacts sensitive information from its logs according to
references/output-redaction.yml.
Audit Metadata