fix(ci): expose jobs cancelled by timeouts

[konard]

Summary

• add a terminal Pipeline Status job that observes every other CI/CD job
• fail aggregate runs for ordinary job failures and for cancelled jobs on main
• preserve warning-only handling for superseded runs on non-default refs
• prevent read-only jobs from being superseded on main, making cancellation classification reliable
• add structural and behavioural regression tests plus a changelog fragment

Reproduction

A job terminated by its own timeout-minutes receives the GitHub Actions conclusion cancelled, not failure. Without an aggregate observer, a run containing that timeout can remain grey and absent from failed-run filters.

Verification

• pytest — 71 passed
• ruff check .
• ruff format --check .
• mypy src
• python scripts/check_file_size.py
• workflow YAML parse and bash -n scripts/check-pipeline-status.sh

The behavioural test covers successful/skipped jobs, an ordinary failure, a cancellation on main, and a cancellation on another ref. A structural test verifies the gate continues to needs: every other job as the workflow evolves.

Fixes #50

[konard]

Working session summary

Implemented and finalized PR #53.

Key changes:

• Added terminal Pipeline Status gate covering every workflow job.
• Main-branch cancellations now fail visibly; PR supersedes remain warnings.
• Added structural and behavioral regression tests.
• Added changelog fragment and removed WIP metadata.
• Committed and pushed as f8e0eda.

Verification:

• 71 tests passed.
• Ruff, formatting, mypy, file-size, YAML, and shell syntax checks passed.
• Fresh CI/CD, Security, CodeQL, dependency review, and link checks all passed.
• PR is ready for review; working tree is clean.

---

This summary was automatically extracted from the AI working session output.

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Model: GPT-5.6 Sol
• Provider: OpenAI
• Public pricing estimate: $3.578604

📊 Context and tokens usage:

• 110.2K / 200K (55%) input tokens, 10.8K / 128K (8%) output tokens

Total: (110.2K + 2.0M cached) input tokens, 10.8K output tokens, $3.578604 cost

🤖 Models used:

• Tool: OpenAI Codex
• Requested: gpt-5.6-sol
• Thinking level: low (~7999 tokens)
• Model: GPT 5.6 Sol (gpt-5.6-sol)

📎 Log file uploaded as Gist (1210KB)

• View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

🎉 Auto-merged

This pull request has been automatically merged by hive-mind.

• All CI checks have passed

---

Auto-merged by hive-mind with --auto-merge flag