docs: Add verification examples to skill documentation

[marcfargas]

Problem

Our skill documentation shows how to perform operations but often lacks verification steps. This gap was exposed when investigating follower notifications (#19) - we had examples for message_subscribe and message_post but no verification that followers were actually added or notifications created.

Example of the gap:

// We documented this:
await client.call('res.partner', 'message_subscribe', [[partnerId]], {
  partner_ids: [followerPartnerId]
});
// But not this verification:
const followers = await client.searchRead('mail.followers', [
  ['res_model', '=', 'res.partner'],
  ['res_id', '=', partnerId],
  ['partner_id', '=', followerPartnerId]
], { fields: ['id'] });
assert(followers.length > 0, 'Follower should be added');

Proposal

Update skill documentation to include verification patterns after operations:

1. CRUD operations - Verify with read/search after create/write/unlink
2. Method calls - Verify side effects (followers added, notifications created, etc.)
3. State changes - Verify the expected state after actions

Benefits

• Agents can self-verify their work
• Catches silent failures (like kwargs being ignored)
• Reinforces testing culture
• Documents expected behavior more completely

Files to Update

• mail/chatter.md - Add verification after message posting
• mail/discuss.md - Add verification for channel operations
• mail/activities.md - Add verification for activity creation
• base/crud.md - Add verification patterns section
• Other module skills as appropriate

Related

Discovered while investigating notification behavior - the root cause was issue #19 (execute vs execute_kw), but better verification examples would have caught this sooner.