The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs bash scripts (e.g., manage-sprint-state.sh, move-sprint-issues.sh) to automate Jira sprint workflows. These scripts are invoked via the shell to interact with the Jira Agile API. \n- [CREDENTIALS_UNSAFE]: The skill manages Jira and Bitbucket API tokens securely. It uses Authorization: Basic headers with Base64 encoding and implements an extensive redaction system in defer-mutation.js and jira-sprint-lib.sh. This system masks credentials and high-entropy strings in the journal and logs, replacing them with variable names or placeholders to prevent exposure. \n- [DYNAMIC_EXECUTION]: Internal policy enforcement relies on dynamic execution patterns, such as spawnSync in defer-mutation.js to run the platform resolver and eval in resolve-platform.sh for cross-shell variable expansion. These operations are strictly limited to internal configuration management and security gating. \n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted issue summaries and keys from Jira. This attack surface is mitigated by the 'access gate' mechanism in jira-sprint-lib.sh, which ensures that any state-changing operations are reviewed or deferred unless a 'full' access mode is explicitly configured by the user.

jira-sprint-manager