CIE-Unified

2026-03-31_task04_branch_a_validation_reporting.md

  1# Task 04: Branch A Validation and Reporting
  2
  3## Title
  4
  5Task 04: unified validation + reporting against locked spec
  6
  7## Direct Prompt
  8
  9Continue Branch A from commit `164c06af63812c69cab32d8f8a6c770b96f38ef6`, keep the implementation independent from other branches, add a small unified validation/reporting entrypoint against the locked docs, generate machine-readable and human-readable reports, and record execution results in-repo.
 10
 11## Suggested Branch Name
 12
 13`branch-a/task04-validation-reporting`
 14
 15## Goal
 16
 17Implement the smallest explicit Branch A validation/reporting layer that checks the locked runtime surface, summarizes smoke/dynamics/sedimentation readiness in one schema, and documents Branch A as ready for review/comparison.
 18
 19## Background
 20
 21This round continues to follow the locked conceptual and engineering constraints in:
 22
 23- `/Users/george/code/CIE-Unified/README.md`
 24- `/Users/george/code/CIE-Unified/LOCKED_IMPLEMENTATION_SPEC.md`
 25
 26Those documents remain locked and define the Branch A requirements for graph-native runtime state, the unified interface, output-to-input feedback, homing, decay, degraded output, sedimentation observability, and comparable reporting.
 27
 28## Involved Repo
 29
 30- `/Users/george/code/CIE-Unified`
 31
 32## Scope
 33
 34- update the Branch A plan for Task 04 completion
 35- create this Task 04 prompt document in-repo
 36- add a stdlib-only `cie.validation` entrypoint for compact controlled scenarios
 37- validate smoke, dynamics, sedimentation, and locked snapshot requirements in a consistent schema
 38- generate a machine-readable JSON report and a human-readable Markdown report
 39- add tests for the validation entrypoint and report schema
 40- append the Task 04 continuation note to the Task 03 record
 41- leave Branch A documented as ready for review/comparison rather than adding new runtime subsystems
 42
 43## Allowed Modifications
 44
 45- `/Users/george/code/CIE-Unified/plans/2026-03-31_branch_a_plan.md`
 46- `/Users/george/code/CIE-Unified/tasks/2026-03-31_task03_branch_a_sedimentation.md`
 47- `/Users/george/code/CIE-Unified/tasks/2026-03-31_task04_branch_a_validation_reporting.md`
 48- `/Users/george/code/CIE-Unified/reports/2026-03-31_task04_branch_a_validation.md`
 49- `/Users/george/code/CIE-Unified/reports/2026-03-31_task04_branch_a_validation.json`
 50- `/Users/george/code/CIE-Unified/cie/__init__.py`
 51- `/Users/george/code/CIE-Unified/cie/runtime.py`
 52- `/Users/george/code/CIE-Unified/cie/validation.py`
 53- `/Users/george/code/CIE-Unified/tests/__init__.py`
 54- `/Users/george/code/CIE-Unified/tests/test_smoke.py`
 55- `/Users/george/code/CIE-Unified/tests/test_dynamics.py`
 56- `/Users/george/code/CIE-Unified/tests/test_sedimentation.py`
 57- `/Users/george/code/CIE-Unified/tests/test_validation.py`
 58
 59## Avoid Modifying
 60
 61- `/Users/george/code/CIE-Unified/README.md`
 62- `/Users/george/code/CIE-Unified/LOCKED_IMPLEMENTATION_SPEC.md`
 63
 64## Must Complete
 65
 66- mark Task 03 complete and Task 04 complete in the Branch A plan
 67- create this Task 04 prompt document in-repo
 68- implement a unified validation/reporting entrypoint for Branch A
 69- validate smoke + dynamics + sedimentation expectations in one consistent report schema
 70- check locked `snapshot_state()` fields and required runtime behaviors
 71- generate `/Users/george/code/CIE-Unified/reports/2026-03-31_task04_branch_a_validation.json`
 72- generate `/Users/george/code/CIE-Unified/reports/2026-03-31_task04_branch_a_validation.md`
 73- run the recommended validation commands
 74- record execution details and remaining review-stage limitations here
 75
 76## Acceptance Criteria
 77
 781. there is a unified validation/reporting entrypoint for Branch A
 792. validation covers smoke + dynamics + sedimentation expectations in a consistent schema
 803. validation checks locked `snapshot_state` fields and required runtime behaviors
 814. a machine-readable JSON report is generated
 825. a human-readable Markdown report is generated
 836. tests pass
 847. Branch A is documented as ready for review/comparison
 85
 86## Evaluation Requirements
 87
 88- use only the Python standard library in new runtime/validation code
 89- keep Branch A graph-native with `(phi, mu, J)` as the canonical state
 90- keep state minimal, parameters few, and observability explicit
 91- avoid `exact_text_map`
 92- avoid MoE-style substitution
 93- avoid latent-vector ontology as the real runtime state
 94- prefer an explicit readable harness over a larger framework
 95
 96## Recommended Validation Commands
 97
 98- `python3 -m unittest discover -s tests -v`
 99- `python3 -m cie.validation`
100
101## Delivery Requirements
102
103- commit on `branch-a/task04-validation-reporting`
104- push the branch to `origin`
105- keep the implementation independent from `branch-b` and other later branches
106- include execution record details for branch/base/backup/files/validation/results/limitations
107- leave Branch A ready for review/comparison rather than feature-expanded
108
109## Execution Record
110
111- actual branch name: `branch-a/task04-validation-reporting`
112- base commit: `164c06af63812c69cab32d8f8a6c770b96f38ef6`
113- backup path used for dirty-worktree handling: `none`
114- files changed:
115  - `/Users/george/code/CIE-Unified/plans/2026-03-31_branch_a_plan.md`
116  - `/Users/george/code/CIE-Unified/tasks/2026-03-31_task03_branch_a_sedimentation.md`
117  - `/Users/george/code/CIE-Unified/tasks/2026-03-31_task04_branch_a_validation_reporting.md`
118  - `/Users/george/code/CIE-Unified/cie/validation.py`
119  - `/Users/george/code/CIE-Unified/tests/test_validation.py`
120  - `/Users/george/code/CIE-Unified/reports/2026-03-31_task04_branch_a_validation.json`
121  - `/Users/george/code/CIE-Unified/reports/2026-03-31_task04_branch_a_validation.md`
122- validation commands:
123  - `python3 -m unittest discover -s tests -v`
124  - `python3 -m cie.validation`
125- concise test summary: `Ran 18 tests; all passed.`
126- concise report summary: `Validation report status PASS; interface, smoke, dynamics, sedimentation, and snapshot sections all passed; Branch A is ready for review/comparison.`
127- remaining known limitations for review stage:
128  - the validation harness is compact and scenario-based rather than a benchmark or long-run stability suite
129  - checks target the locked observable runtime surface, not richer semantic task-performance evaluation
130  - homing and sedimentation remain explicit but heuristic, which is acceptable for this comparison-ready stage