feat: consolidate issue 54 ingestion and promotion gates

[steezkelly]

Summary

Consolidates the open issue #54 implementation stack and the overlapping prerequisite safety fixes from PRs #60, #61, #63, #64, #65, and #66 into one reviewable PR.

This is a consolidation PR because direct merge was unavailable to the contributor account and local integration found a real overlap between #61 and #64 in evolution/skills/evolve_skill.py. Keeping the stack split would require maintainers to resolve that conflict manually after the first overlapping PR lands.

Includes:

• fix: declare reportlab dependency #60: declare reportlab>=4.0 and regression-test report import/install behavior
• fix: fail fast on invalid baseline skills #61: fail fast when baseline skill constraints fail, validating full raw skill frontmatter
• feat: report session source availability #66: report explicit per-source importer availability/candidate counts for all-source dry runs
• feat: write evolution run reports #64: write local machine-readable evolution run reports and failure artifacts
• feat: add benchmark gate for run reports #63: add fail-closed benchmark gate for run reports
• feat: add local-first pr builder #65: add local-first PR builder from run reports

Overlap / conflict resolution

Local integration found one real conflict between #61 and #64:

• fix: fail fast on invalid baseline skills #61 imports ConstraintResult in evolution/skills/evolve_skill.py
• feat: write evolution run reports #64 imports write_run_report in the same import block

The consolidated branch resolves this by keeping both:

from evolution.core.constraints import ConstraintValidator, ConstraintResult
from evolution.core.run_report import write_run_report

No other file conflicts were found in the tested integration stack.

Local verification evidence

GitHub checks were absent on the source PRs: statusCheckRollup length was 0 for #60, #61, #63, #64, #65, and #66 when reviewed with gh pr view.

Local evidence preserved in the working copy at:

/home/steve/repos/hermes-agent-self-evolution/issue-54-pr-stack-local-evidence.md

Commands run in a fresh local venv:

python3 -m venv .venv-review
. .venv-review/bin/activate
pip install -e '.[dev]'
pytest tests/test_generate_report.py tests/skills/test_evolve_skill_constraint_gates.py tests/core/test_external_importer_availability.py tests/core/test_run_report.py tests/core/test_benchmark_gate.py tests/core/test_pr_builder.py -q
pytest -q

Results:

• Targeted stack tests: 21 passed, 11 warnings in 1.33s
• Full suite: 160 passed, 11 warnings in 1.42s

Warnings were DSPy dependency deprecation warnings about the prefix argument in InputField / OutputField; no test failures.

Review / merge guidance

If maintainers prefer the original split PRs instead of this consolidation PR, the tested conceptual merge order is:

1. fix: declare reportlab dependency #60
2. fix: fail fast on invalid baseline skills #61
3. feat: report session source availability #66
4. feat: write evolution run reports #64, with the import conflict resolved as above if fix: fail fast on invalid baseline skills #61 has landed
5. feat: add benchmark gate for run reports #63
6. feat: add local-first pr builder #65

Otherwise, review and merge this PR as the consolidated path, then close the split PRs as superseded.

Closes #54.
Supersedes #60, #61, #63, #64, #65, #66.

[steezkelly]

Closing this PR in favor of consolidated PR #68. Local integration found real helper-block overlap in evolution/skills/evolve_skill.py across the stack, and #68 preserves local test evidence: targeted stack tests 41 passed; full suite 164 passed; GitHub checks were absent on the split PRs. Review #68 instead.