Rewrite README into reviewer-oriented deterministic replay validation front page#145
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request significantly refactors the README.md, shifting the focus toward deterministic replay-survivability validation for compressed operational state in long-horizon AI agents. The update includes new Mermaid diagrams for the validation model, artifact pipeline, and roadmap, alongside a simplified repository map and updated benchmark results. Reviewer feedback recommends ensuring consistent project naming ('CompText V7') in accordance with the style guide and restoring missing environment setup and benchmark testing commands to the local verification section.
| - relational admissibility | ||
|
|
||
| All deterministically validated. | ||
| CompTextv7 does not ask whether a compressed summary sounds good. It asks whether the compressed state can still replay the operational facts required to continue the work. |
There was a problem hiding this comment.
The project name is referred to inconsistently throughout the document (e.g., CompTextv7, Comptextv7). According to the Repository Style Guide (Line 1), the preferred name is CompText V7. Please ensure consistency across the README.
| CompTextv7 does not ask whether a compressed summary sounds good. It asks whether the compressed state can still replay the operational facts required to continue the work. | |
| CompText V7 does not ask whether a compressed summary sounds good. It asks whether the compressed state can still replay the operational facts required to continue the work. |
References
- The project name should be consistently referred to as 'CompText V7' as defined in the repository memory. (link)
| - Profile comparison: fixture-bound aggregates for collapse rate, replay consistency, operational drift, evidence survival, and deterministic failure labels. | ||
| - Sensitivity analysis: bounded variations of `max_context_units`, `max_families`, `max_bursts`, `replay_window_seconds`, `replay_cycles`, and `compression_budget_scale`. | ||
| ```bash | ||
| npm install --no-save --no-package-lock |
There was a problem hiding this comment.
The Python dependency installation step (python -m pip install -e '.[test]') has been removed in this version. Since the local verification process relies on pytest, this step is necessary for users setting up the environment from scratch.
| npm install --no-save --no-package-lock | |
| python -m pip install -e '.[test]' | |
| npm install --no-save --no-package-lock |
| pytest tests/test_failure_taxonomy.py -q | ||
| pytest tests/test_multi_family_admissibility_artifact.py -q | ||
| pytest tests/test_multi_family_svg_renderer.py -q |
There was a problem hiding this comment.
Several benchmark tests (e.g., test_paper_replay_bench.py, test_agent_trace_replay.py) are missing from the local verification instructions. These should be included to allow users to verify the results presented in the 'Current fixture-bound signal' table.
| pytest tests/test_failure_taxonomy.py -q | |
| pytest tests/test_multi_family_admissibility_artifact.py -q | |
| pytest tests/test_multi_family_svg_renderer.py -q | |
| pytest tests/test_failure_taxonomy.py -q | |
| pytest tests/test_multi_family_admissibility_artifact.py -q | |
| pytest tests/test_multi_family_svg_renderer.py -q | |
| pytest tests/test_paper_replay_bench.py tests/test_agent_trace_replay.py -q |
1 participant
README-only rewrite preserving the existing logo/header while adding reviewer-oriented structure, deterministic positioning, GitHub-native Mermaid diagrams, fixture-family coverage, failure taxonomy mapping, CI gate visualization, artifact evidence pipeline, verification commands, roadmap, and conservative fixture-bound framing.
Scope intentionally limited to README.md only.