Scan agent-derived code and text for secrets

[eshulman2]

Summary

• compose repository secret detection into Add a fail-closed pre-push validation gate in GitOperations #263's common OutputValidator pre-push gate
• scan changed repository files using the trusted base and trusted-base .secrets.baseline
• add an explicit provenance-aware API for workflow nodes publishing agent-derived text
• scan agent-derived PR descriptions, Q&A responses, review objections, and post-merge summaries at their workflow origin
• keep Jira/GitHub clients as trusted transport components rather than global policy interceptors
• expose only redacted detector, path/source, and line information

Stack

This PR is intentionally based on issue-263-forge-native; review and merge #263 first, then retarget/rebase this PR to main.

Test plan

• 25 focused repository/output/secret tests pass
• 35 affected workflow-node tests pass
• Ruff passes on changed implementation and tests

Closes #77

Depends on #263