Repository navigation
Fix ISS-014: add model/task fitness check - #102
Merged
Merged
Conversation
New py_mono/skill/fitness.py: check_model_fitness(skill, provider, model) reads ISS-013's telemetry log and, once at least 3 recorded runs exist for the exact (skill, provider, model) combination, warns if at least half of the most recent 5 matching runs failed. Returns None (no opinion) below the minimum sample size, avoiding a false-positive warning from a single bad run. Hooked into run_skill_safe (py_mono/skill/approval.py): a returned warning is prepended to a successful result; a failed run attaches no warning, since its own exception is already the signal. Directly productizes the ISS-009/ISS-011 lesson (thinking models reasoning verbosely/unreliably on structured, template-filling tasks) as an evidence-based check driven by this repo's own recorded history, not a hardcoded list of "bad" models. Adds tests/test_skill_fitness.py (6 tests) and 3 new tests in tests/test_skill_approval.py. This is the last item in Milestone 6 - see docs/ROADMAP_PLAN.md and specs/012-add-model-task-fitness-check/.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
py_mono/skill/fitness.py:check_model_fitness(skill, provider, model)readsISS-013's telemetry log and, once at least 3 recorded runs exist for the exact combination, warns if at least half of the most recent 5 matching runs failed. No opinion (returnsNone) below the minimum sample size — avoids a false-positive from a single bad run.run_skill_safe(py_mono/skill/approval.py): a warning is prepended to a successful result; a failed run attaches no warning.ISS-009/ISS-011lesson (thinking models reasoning verbosely/unreliably on structured tasks) as an evidence-based check, not a hardcoded model list.ISS-013(Fix ISS-013: add lightweight per-skill-run telemetry log #101) — merged into this branch for validation, same pattern as Fix ISS-012: add minimal CI (pytest + compileall on every PR) #100/Fix ISS-005: root-cause three pre-existing test failures #96.ISS-014indocs/ISSUES.md; SDD trail inspecs/012-add-model-task-fitness-check/.docs/ISSUES.md/docs/ROADMAP_PLAN.md— expect merge conflicts on those two files depending on merge order. Once all seven are onmain, the Milestone 6 status line and Milestone Index indocs/ROADMAP_PLAN.mdshould get a final pass to say "done."Test plan
pytest tests/test_skill_fitness.py tests/test_skill_approval.py -v— 16 passedISS-005failures tracked/fixed separately in Fix ISS-005: root-cause three pre-existing test failures #96, unrelated to this change)python -m compileall -q py_mono skills— clean