[KatalKavya96]

Summary

• Fix missing vetted-ops caller permissions for the three security issue-import workflows by documenting their exact minimum grants in Setup configuration.
• Extend setup reconcile to detect missing callers or operations, propose a minimal policy diff, and require explicit confirmation before applying changes.
• Add a reconciliation evaluation fixture covering all three import callers, extend the evaluation output contract, and update case counts.

Type of change

• Skill change (.claude/skills/<name>/) — eval fixtures updated below
• Tool / bridge contract (tools/<system>/*.md)
• Python package (tools/*/ with pyproject.toml)
• Groovy reference impl
• Cross-cutting (RFC, AGENTS.md, sandbox, privacy-LLM)
• Documentation (docs/, README.md, CONTRIBUTING.md)
• Project template (projects/_template/)
• CI / dev loop (prek, workflows, validators)
• Other: Setup skill instructions and reconciliation evaluation contract

Test plan

• prek run --all-files passes
• Workspace pytest, Ruff check, Ruff format, and mypy passed
• Groovy bridges — not changed
• Model-based step-reconcile evaluation: 4 passed, 0 failed
• New case-4-import-policy-grants regression fixture added

RFC-AI-0004 compliance

• HITL — policy edits require explicit confirmation
• Sandbox — no new unrestricted access
• Vendor neutrality — existing adopter policy conventions preserved
• Conversational + correctable — missing grants presented individually
• Write-access discipline — no automatic permission changes; existing grants preserved
• Privacy LLM — not affected

Linked issues

Closes #1731

Notes for reviewers

The current import workflows require only:

• security-issue-import: rollup-append
• security-issue-import-from-md: rollup-append
• security-issue-import-from-pr: pr-view-with-body, rollup-append

Although the issue mentions body-field-set, no import workflow currently invokes it, so this PR intentionally does not grant it.

The vetted-ops dispatcher and tracker procedures are unchanged. Policy updates remain subject to explicit user confirmation.

AI assistance was used for investigation, patch drafting, and review. Changes were applied and validated locally by the contributor.

[KatalKavya96]

Hi @potiuk ,

The fix for #1731 is ready for review.

What changed:

• Documented the minimum vetted-ops permissions for all three security issue-import callers.
• Extended setup reconcile to detect missing callers and operations in existing policies.
• Preserved explicit user confirmation before policy edits, existing permissions, and unrelated caller configurations.
• Added a regression fixture covering all three missing-grant scenarios.

Validation completed:

• prek run --all-files — all hooks passed.
• Model-based step-reconcile evaluation — 4/4 passed, 0 failed.

One clarification: although the issue mentions body-field-set, none of the current import workflows invokes it, so it was intentionally excluded to maintain least-privilege access.
Thanks!