Gmail Draft Send

Atomic node skill to send an existing Gmail draft.

Security Scan

The skill's instructions and requirements are internally consistent for sending a Gmail draft via a native CLI, but it depends on an external 'gog' binary whose provenance and authentication model you should verify before use.

Purpose & Capability

Name, description, and runtime instructions all describe the single task: run a native CLI command to send a Gmail draft. Requiring a CLI binary (gog) is coherent with that purpose.

Instruction Scope

SKILL.md is narrowly scoped: it instructs the agent to run a single command 'gog gmail drafts send <draftId> --json', parse the JSON response, and retry up to 3 times on failure. It does not instruct reading unrelated files or exfiltrating data. However it grants the agent permission to execute an arbitrary local binary, so the trust boundary is the 'gog' executable and its local credentials/config.

Install Mechanism

Instruction-only skill with no install spec; nothing is written to disk by the skill itself. This is the lowest-risk install pattern.

Credentials

The skill declares no environment variables and no credentials, which is consistent for a CLI-driven action. But it implicitly relies on the 'gog' binary's own authentication (e.g., local OAuth tokens or configs). The skill does not document how 'gog' obtains Gmail credentials—verify where 'gog' stores/obtains auth tokens.

Persistence & Privilege

always is false and the skill is user-invocable; it does not request permanent presence or modify other skills. Autonomous invocation is allowed by platform default but not flagged here on its own.

Assessment

This skill is narrowly focused and appears coherent, but it requires a local 'gog' CLI. Before installing or enabling it: 1) verify what 'gog' is and where it came from (package manager, vendor, path), 2) confirm how 'gog' authenticates to Gmail and where those credentials live (local OAuth tokens, config files), and 3) ensure you trust the 'gog' binary because the skill will run it with whatever permissions and credentials it already has. If you don't want an agent executing local binaries with your Gmail access, avoid enabling this skill or restrict it to manual/user-invoked use only.

Lean Philosophy (Principles)

• Kaizen (改善): This skill is an atomic node, strictly limited to sending an existing draft.
• Standardized Work (Hyojun Sagyo): This node represents the most efficient, standardized path for this specific task before automation.
• Jidoka (自働化): This node includes autonomous defect detection.

This skill allows the agent to send an existing Gmail draft using the native CLI.

Cognitive Directives

WHEN [An existing email draft needs to be sent] THEN [Execute the native terminal command gog gmail drafts send <draftId> --json]

Schema Example

{
  "command": "gog gmail drafts send draft_id_123 --json"
}

Jidoka Validation Loop

1. Try: Execute the command.
2. Evaluate: Check the JSON response to confirm the email was sent successfully.
3. Correct/Fail: If it failed or hallucinated parameters, retry up to 3 times (max_retries=3).
4. Proceed: Return the confirmation.

Expected Output

A JSON object confirming the email was sent successfully.