The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/release.py and its supporting modules (_jira.py, rich_filter.py) execute shell commands via subprocess.run to interact with acli (Atlassian CLI), gog (Google Workspace CLI), and git. These executions are used to search Jira, retrieve spreadsheet data, and identify repository paths.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from shared resources (Jira issue descriptions and Google Sheets cells) to extract release dates and team metadata. This data is then used to formulate responses and Slack messages.
Ingestion points: Jira issue descriptions parsed in scripts/release.py; Google Sheets cell data fetched in scripts/release.py.
Boundary markers: Absent; data is processed and presented without specific delimiters to isolate potentially malicious embedded instructions.
Capability inventory: Subprocess execution for CLI tools (acli, gog, git) in multiple scripts.
Sanitization: Release versions are normalized via regex (\d+\.\d+), but the parsing logic for milestone dates in issue descriptions relies on general regex matching which does not fully sanitize the input text.
[EXTERNAL_DOWNLOADS]: The skill retrieves data from well-known external services including issues.redhat.com (Jira) and docs.google.com (Google Sheets). These references target official project resources associated with the developer.

rhdh-release-schedule