We split this skill into the rules it gives an agent

8 rules. About 0 are common, 2 are distinctive, and 6 are very specific to this project.

• ≈0

  Common · estimate

  Many files like this have a rule of this kind.

• ≈2

  Distinctive · estimate

  This author's own choices. Worth a look.

• 6

  Very specific

  Tied to their project's names, files or tools. Probably skip.

• Common
• Distinctive
• Very specific

1. Distinctive:

   Use this when the user wants Samantha to act through an external connector beyond read-only setup/backfill: sending messages, replying to …

   line 13
2. Distinctive:

   Do not treat a connected account as permission to act.

   line 17
3. Very specific:

   Prefer narrow constraints: allowed_targets / allowed_recipients and max_per_day are safer than broad grants.

   line 19
4. Very specific:

   Keep destructive actions, especially delete, out of autonomous policy unless the user is exceptionally explicit and the constraints are …

   line 21
5. Very specific:

   Call connector_action_policy_status to inspect existing grants.

   line 25
6. Very specific:

   If granting, summarize the proposed scope in normal language before calling grant_connector_action_policy.

   line 26
7. Very specific:

   Use requires_per_action_approval: false only when the user asks for preauthorization rather than case-by-case approval.

   line 27
8. Very specific:

   Use revoke_connector_action_policy when the user asks to remove or narrow a grant.

   line 28

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.