The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the vendor's CLI tool, postplus, to perform research tasks, such as postplus research collect for data gathering and postplus research schema for structure validation.
[EXTERNAL_DOWNLOADS]: The skill fetches public data from X via the vendor's hosted collection services during the execution of research commands.
[PROMPT_INJECTION]: The skill processes untrusted public content from X, which presents a surface for indirect prompt injection.
Ingestion points: Data enters the context from X posts, profiles, and replies via the result.json file generated by the CLI (referenced in SKILL.md).
Boundary markers: The skill employs a shared-contract.md which defines explicit data schemas and collection bounds (e.g., maxItems: 20) to limit the volume of untrusted data.
Capability inventory: The skill can execute shell commands via the postplus CLI and perform file system operations to read/write request and result JSON files.
Sanitization: Instructions in references/shared-contract.md mandate separating raw post/profile facts from analysis and explicitly state to report sparse or low-signal output as gaps rather than processing it blindly.
[NO_CODE]: The skill package contains only markdown instructions and configuration files; it does not include any standalone scripts or binary executables.

x-research