The Agent Skills Directory

[DATA_EXFILTRATION]: The skill provides instructions for reading public key files from the ~/.ssh/ directory and uploading them to GitLab via the glab CLI. While this involves interacting with a sensitive directory and performing network operations, it is the primary intended purpose of the skill. The instructions include a specific 'Security Warning' section and verification steps (e.g., checking for the .pub extension and inspecting file headers) to ensure that private keys are not accidentally exposed.
[INDIRECT_PROMPT_INJECTION]: The skill utilizes the output of glab ssh-key list and glab ssh-key get, which ingest data from the GitLab API. This data (such as SSH key titles) could theoretically be controlled by an external party and contain malicious instructions.
Ingestion points: Command output from glab ssh-key list and glab ssh-key get (SKILL.md, references/commands.md).
Boundary markers: No specific delimiters or instructions to ignore embedded content are provided.
Capability inventory: The skill possesses the capability to read files from the filesystem (~/.ssh/) and execute network-enabled commands via the glab tool.
Sanitization: No sanitization or validation of the structured or text output from the CLI is described.
[COMMAND_EXECUTION]: The skill instructs the agent on how to execute several glab subcommands. These are standard administrative operations for managing account credentials on a well-known service (GitLab).

glab-ssh-key