fix(mcp): make local-endpoint budgets configurable

[NullSense]

Summary

• tool_answer/config.py: _SYNTHESIS_MAX_TOKENS (1024, hardcoded) now reads REPOWISE_SYNTHESIS_MAX_TOKENS. A reasoning model spends its whole budget on hidden thinking before emitting an answer token, so the call comes back empty at the old cap; measured against nanbeige, 1024 tokens returned nothing, 8192 produced a correct, cited answer.
• _answer_pipeline.py / _helpers.py: _EMBED_TIMEOUT_S (8.0, hardcoded) now reads REPOWISE_EMBED_TIMEOUT_S via a new embed_timeout_s(), mirroring the sibling vector_search_timeout_s(). A cold-loading local embedding model blows the old 8s budget; get_answer still exits 0 with the semantic leg silently dropped, visible only at WARNING.
• One PR because it's one root cause: constants tuned for a warm remote service, with no override, that degrade silently on a local one. Both follow the existing pattern for REPOWISE_EMBEDDING_TIMEOUT / REPOWISE_VECTOR_SEARCH_TIMEOUT_S: an unparseable or non-positive override warns and keeps the default.

Related Issues

None.

Test Plan

• uv run pytest tests/unit/ -q — 14242 passed, 12 skipped, 2 xfailed. 2 failures, both pre-existing on unmodified main and unrelated to this change: a git-rebase test that depends on local git config, and a timezone-sensitive episode-dating test.
• uv run ruff check .
• uv run repowise risk main..HEAD:

  Change risk for main..HEAD: touches files that have broken before · 90th percentile of this repo's fix-bearing files
    fix(mcp): make local-endpoint budgets configurable
    +269 / -12 lines · 7 files · 3 dirs · 2 subsystems · entropy 2.19 · author exp 0  (fix)
  
  File                                                                       Lines   Prior fixes
  packages/server/src/repowise/server/mcp_server/tool_answer/config.py         36           11.7
  packages/server/src/repowise/server/mcp_server/_helpers.py                   36            6.4
  tests/unit/server/mcp/test_answer_synthesis_timeout.py                       47            5.8
  packages/server/src/repowise/server/mcp_server/_answer_pipeline.py           19            4.9
  packages/server/src/repowise/server/mcp_server/tool_answer/synthesis.py       6            2.8
  
  Diff shape: Typical · 57th percentile of recent commits by size and spread
    About as risky as a typical commit in this repo.
  

  This flags real history: these are exactly the answer-pipeline files this fix targets, and they carry a documented run of prior fixes (config.py alone shows a recency-weighted 11.7). That matches the PR's own premise — this is a bug-prone area — rather than indicating a new problem in this diff; the full-suite result above is the direct check on that.
  (impacted-tests / health --file need a completed repowise init against this clone; that indexing run did not finish in time and is omitted rather than faked.)

Checklist

• My code follows the project's code style
• I have added tests for new functionality
• All existing tests still pass
• I have updated documentation if needed

---

Written with AI assistance; measurements and tests were run and verified locally.

[repowise-bot]

✅ Health of changed files: 3.2 → 3.6 (+0.3)
_{⚠️ Change risk: moderate, riskier than 54% of this repo's commits.}

📋 At a glance
5 hotspots touched · 2 co-change pairs left out · 5 files with recent fix history.

Files & modules (2)

• packages (4 files)
  • .../mcp_server/_answer_pipeline.py
  • .../tool_answer/config.py
  • .../mcp_server/_helpers.py
  • .../tool_answer/synthesis.py
• tests (1 file)
  • .../mcp/test_answer_synthesis_timeout.py

✅ Health gate: passed

📌 Before you merge

• Run .../mcp/test_answer_agreement.py, .../mcp/test_answer_confidence_reasons.py, .../mcp/test_answer_decision_rows.py, .../mcp/test_decision_hygiene.py (+8 more): they import the changed files
• .../tool_answer/answer.py changed together with .../tool_answer/config.py in 23 past commits and isn't in this PR
• .../tool_answer/symbols.py changed together with .../tool_answer/config.py in 10 past commits and isn't in this PR

🔎 More signals (3)

🗺️ Change map

flowchart LR
  subgraph PR ["Changed in this PR (4 with dependents)"]
    f_packages_server_src_repowise_server_mcp_server__answer_pipeline_py[".../mcp_server/_answer_pipeline.py 🔥"]:::changed
    f_packages_server_src_repowise_server_mcp_server__helpers_py[".../mcp_server/_helpers.py 🔥"]:::changed
    f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py[".../tool_answer/config.py 🔥"]:::changed
    f_packages_server_src_repowise_server_mcp_server_tool_answer_synthesis_py[".../tool_answer/synthesis.py 🔥"]:::changed
  end
  f_packages_server_src_repowise_server_mcp_server__neighbor_rerank_py[".../mcp_server/_neighbor_rerank.py"]
  f_packages_server_src_repowise_server_mcp_server__answer_pipeline_py --> f_packages_server_src_repowise_server_mcp_server__neighbor_rerank_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_answer_py[".../tool_answer/answer.py"]
  f_packages_server_src_repowise_server_mcp_server__answer_pipeline_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_answer_py
  f_packages_server_src_repowise_server_mcp_server_tool_search_py[".../mcp_server/tool_search.py"]
  f_packages_server_src_repowise_server_mcp_server__answer_pipeline_py --> f_packages_server_src_repowise_server_mcp_server_tool_search_py
  f_packages_server_src_repowise_server_mcp_server__prose_symbols_py[".../mcp_server/_prose_symbols.py"]
  f_packages_server_src_repowise_server_mcp_server__helpers_py --> f_packages_server_src_repowise_server_mcp_server__prose_symbols_py
  f_packages_server_src_repowise_server_mcp_server__helpers_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_answer_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_cache_py[".../tool_answer/cache.py"]
  f_packages_server_src_repowise_server_mcp_server__helpers_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_cache_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_evidence_py[".../tool_answer/evidence.py"]
  f_packages_server_src_repowise_server_mcp_server__helpers_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_evidence_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_answer_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_bodies_py[".../tool_answer/bodies.py"]
  f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_bodies_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_cache_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_confidence_py[".../tool_answer/confidence.py"]
  f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_confidence_py
  f_packages_server_src_repowise_server_mcp_server__server_py[".../mcp_server/_server.py"]
  f_packages_server_src_repowise_server_mcp_server_tool_answer_synthesis_py --> f_packages_server_src_repowise_server_mcp_server__server_py
  f_packages_server_src_repowise_server_mcp_server_tool_answer_synthesis_py --> f_packages_server_src_repowise_server_mcp_server_tool_answer_answer_py
  more(["+24 more dependents"])
  PR --> more
  w_packages_server_src_repowise_server_mcp_server_tool_answer_answer_py(["⚠️ .../tool_answer/answer.py changed together 23×, not in PR"]):::warn
  f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py -.- w_packages_server_src_repowise_server_mcp_server_tool_answer_answer_py
  w_packages_server_src_repowise_server_mcp_server_tool_answer_symbols_py(["⚠️ .../tool_answer/symbols.py changed together 10×, not in PR"]):::warn
  f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py -.- w_packages_server_src_repowise_server_mcp_server_tool_answer_symbols_py
  t_tests_unit_server_mcp_test_answer_agreement_py(["✅ .../mcp/test_answer_agreement.py"]):::guard
  t_tests_unit_server_mcp_test_answer_agreement_py -.-> f_packages_server_src_repowise_server_mcp_server__answer_pipeline_py
  t_tests_unit_server_mcp_test_decision_hygiene_py(["✅ .../mcp/test_decision_hygiene.py"]):::guard
  t_tests_unit_server_mcp_test_decision_hygiene_py -.-> f_packages_server_src_repowise_server_mcp_server__helpers_py
  t_tests_unit_server_mcp_test_answer_cache_py(["✅ .../mcp/test_answer_cache.py"]):::guard
  t_tests_unit_server_mcp_test_answer_cache_py -.-> f_packages_server_src_repowise_server_mcp_server_tool_answer_config_py
  classDef changed fill:#dbeafe,stroke:#1d4ed8,color:#1e3a5f
  classDef warn fill:#fef3c7,stroke:#b45309,color:#78350f
  classDef guard fill:#dcfce7,stroke:#15803d,color:#14532d

Loading

_{Solid arrows: code that imports the changed files (33 direct dependents, from the last indexed snapshot). Dashed: history/tests.}

🔥 Hotspots touched (5)

• .../mcp_server/_answer_pipeline.py: 15 commits/90d, 13 dependents · primary owner: Raghav Chamadiya (82%)
• .../mcp/test_answer_synthesis_timeout.py: 7 commits/90d, 0 dependents · primary owner: Raghav Chamadiya (99%)
• .../tool_answer/config.py: 27 commits/90d, 21 dependents · primary owner: Swati Ahuja (59%)

2 more

• .../mcp_server/_helpers.py: 11 commits/90d, 33 dependents · primary owner: Raghav Chamadiya (91%)
• .../tool_answer/synthesis.py: 6 commits/90d, 6 dependents · primary owner: Swati Ahuja (51%)

🔗 Hidden coupling (1 file)

• .../tool_answer/config.py co-changes with these files (not in this PR):
  • .../tool_answer/answer.py (23×, 🔴 strong)
  • .../tool_answer/symbols.py (10×, 🟡 notable)

👀 Suggested reviewers @RaghavChamadiya

---

📊 See the full report for this PR
_{Your repo map with this PR's blast radius lit up, every caller of the contracts it changes, and health before and after. No sign-in. · ⭐ Star Repowise · 📥 Install bot · Silence on a single PR with [skip repowise] in the title · Per-repo toggle on repowise.dev/settings?tab=bot · Updated 2026-08-22 15:46 UTC}

[Ayush7614]

Verified: pytest tests/unit/server/mcp/test_answer_synthesis_timeout.py tests/unit/server/mcp/test_embed_timeout.py -q 76 passed. Makes two hardcoded budgets configurable (REPOWISE_SYNTHESIS_MAX_TOKENS for reasoning models spending budget on hidden thinking, REPOWISE_EMBED_TIMEOUT_S via embed_timeout_s() mirroring vector_search_timeout_s()) with validate-and-warn fallback, correct 120s cap, and updated degraded message advertising the env var. Hotspot history (config.py 11.7 fixes) matches the bug-prone area being fixed — not a new risk. LGTM.