fix(2115): show usage in working session summaries

[konard]

Problem

Automated Working session summary comments contained only the AI's final text.
Cost and token/context-budget data was calculated later and supplied exclusively
to the attached log comment. The same omission affected top-level, watch, and
auto-restart-until-mergeable working sessions.

The preserved incident log proves this was not missing telemetry: Codex emitted
usage and Hive Mind parsed 511,606 input, 23,370,240 cache-read, and 63,787
output tokens before publishing a summary without those facts.

Solution

• build the top-level session budget data before summary publication and reuse
  it for the subsequent log comment;
• render summary cost through the existing buildCostInfoString() component;
• render context fill and token totals through the existing
  buildBudgetStatsString() component;
• pass usage details through top-level, watch, and
  auto-restart-until-mergeable summary paths;
• preserve summaries without an empty metadata block when no usage is observed;
• add patch release metadata and focused regression coverage.

Reproduction and verification

Before implementation, the regression test failed because no working-session
usage formatter existed.

Final local gates:

• node tests/test-issue-2115-working-session-summary-usage.mjs;
• npm test — all 355 selected test files passed;
• npm run lint;
• npm run format:check;
• bash scripts/check-mjs-syntax.sh;
• bash scripts/check-file-line-limits.sh;
• git diff --check.

Preserved evidence

The issue 2115 case study
contains the complete compressed 26,012-line session log, issue and PR comment
metadata, evidence hashes, timeline, requirement inventory, root-cause analysis,
alternatives, and primary-source external research.

Fixes #2115

[konard]

Working session summary

Implemented and finalized issue #2115.

• Working-session summaries now display cost and token/context-budget usage.
• Fixed top-level, watch, and auto-restart flows.
• Added regression coverage and a patch changeset.
• Preserved the complete incident log and deep case study.
• All 355 local test files and static checks passed.
• Fresh CI passed on exact commit e8cab597.
• PR is ready for review, merge-clean, and the working tree is clean.

PR: #2116

---

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: $25.426107

📊 Context and tokens usage:

• 250.3K / 200K (125%) input tokens, 23.0K / 128K (18%) output tokens

Total: (250.3K + 21.9M cached) input tokens, 23.0K output tokens, $25.426107 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 (4856KB)

• View complete solution draft log

---

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

[konard]

✅ Ready to merge

This pull request is now ready to be merged:

• All CI checks have passed
• No merge conflicts
• No pending changes

---

Monitored by hive-mind with --auto-restart-until-mergeable flag