[adriannoes]

This PR stacks on #850. After #850 merges, retarget it to dev, then close and reopen it to start full CI.

Summary

• Automations and AI agents choose each part with the process-design selection table before any AI consent question.
• An explicit AI request or an accepted AI line supplies consent. Without either, the skill asks. A later "no AI" removes AI parts.
• Agent lifecycle instructions use the same five off cases as automations. AI automation creation uses the chosen state and reads the saved rule back.
• Agent acceptance compares each conditional outcome with stored card fields and phase. Cleanup waits for a terminal execution status.
• Intelligence and impact use the same selection table. Intelligence also uses the consent rule. Its inactive example describes a pipe with cards.

Test plan

• Clean combined stack: 75 focused tests passed. Linters passed for 19 skills and 57 references.
• Plugin packaging and Claude validation passed. Claude loaded the catalog, and the skills CLI listed all 19 skills.
• Earlier MCP probes cover the unchanged simulation and log recipes. The latest documentation received an offline consumer review.
• Earlier MCP log QA read three completed executions through cursor pagination. Earlier acceptance captured card API results before cleanup. The later MCP run covered logs only.
• New live run of the public card-read recipe: not run.
• The CLI help contains the flags used by the log and card-read recipes.
• Merge simulations against current dev and feat(skills): route whole-process builds in pipefy-building #851: no conflict.
• DCO and skills checks for the latest commit passed.
• Full CI after retargeting to dev.
• Cursor-specific smoke: not run.
• Autonomous run of the final instructions: not run. The new consumer assessment was offline.

Docs / skills

• docs/parity.md updated when MCP ↔ CLI coverage changed (no coverage change)
• Affected skills in plugins/pipefy/skills/ updated in this PR (or a paired PR)
• No docs/skills update needed

Legal / contributions

• Commits include DCO sign-off (git commit -s)
• Regulated-domain skills include COMPLIANCE.md when applicable (none added)