The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute various development and validation tools as part of the Electron build and update lifecycle. This includes running the skill's validation script python3 scripts/validate_skill.py mentioned in README.md, as well as project-specific commands for linting, testing, packaging, signing, and notarization specified in SKILL.md Step 4.
[EXTERNAL_DOWNLOADS]: The skill's verification workflow involves fetching external assets. Evidence in SKILL.md Steps 2 and 4 describes downloading appcast manifest files and .delta update packages from the application's configured release sources for integrity checks.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as update feeds and release pages, which constitutes an attack surface for indirect prompt injection.
Ingestion points: Data is ingested from public release pages and appcast response bodies during the deliverable validation phase in Step 4 of SKILL.md.
Boundary markers: No explicit text delimiters or boundary markers are defined for the external feed content processing.
Capability inventory: The skill possesses capabilities for shell command execution and local file system management during cleanup and restart sequences (SKILL.md Steps 3 and 4).
Sanitization: The workflow relies on cryptographic verification of signatures, digests, and metadata (Step 2) to ensure the integrity of ingested update data.

sgc-electron-delta-updater