The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: Downloads and executes an installation script from https://native.jam.dev/install using curl | bash to set up the CLI environment when the jam command is not found.
[EXTERNAL_DOWNLOADS]: Fetches tools and updates from the vendor's official domain (jam.dev) to facilitate recording analysis and fix verification.
[COMMAND_EXECUTION]: Executes shell commands including jam, ffmpeg, and npx playwright to collect diagnostic data, record videos, and run tests to verify code changes.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external recordings that may contain untrusted content such as reporter notes, console logs, and transcripts.
Ingestion points: Data is retrieved via tools like getDetails, getConsoleLogs, getNetworkRequests, and getVideoTranscript in SKILL.md and references/cli.md.
Boundary markers: There are no explicit instructions to wrap this external content in delimiters or to ignore embedded instructions.
Capability inventory: The agent has the ability to execute shell commands, modify source code, and create comments on external Jam recordings.
Sanitization: No explicit sanitization or filtering logic is provided for the text content extracted from the recordings.

jam-fix