Causal Memory Layer: value, movement, and grant evidence.
A visual landing dashboard for showing why CML matters, what evidence already exists, and how the project moves toward a credible $75k–$100k grant plan.
Evidence momentum
Movement signal: CML now has grant evidence, benchmark baseline, Docker demo, external validation protocol, technical report outline, and a $75k–$100k execution plan.
People are turning CML from artifact into community.
This board recognizes concrete project value: reproducibility, research clarity, documentation, reviewability, and trust signals. The score is a qualitative impact signal, not a competition.
Turned CML into a more reproducible and explainable research artifact: demo path, causal-invalidity taxonomy, agentic workflow boundaries, and strategic integration hypotheses.
Added plain-language explanations for audit findings, making CML easier to understand for new contributors, reviewers, and non-core readers.
Helped surface the importance of containerized local setup. Even where later work superseded the PR, the direction strengthened the demo/reproducibility path.
Proposed API smoke-test coverage for health, audit, and CTAG decode paths. The contribution is useful but still needs fixes before merge.
Where the project value is visible now
The strongest current value is in grant-facing clarity, core audit semantics, and reproducible demos. The next jump comes from external validation and technical reporting.
Causal validity primitive
CML distinguishes operational success from causal legitimacy: what happened is not enough; the system must show why an action was allowed.
Benchmark baseline
Current deterministic benchmark baseline: 6/6 matched cases with explicit audit findings and reproducible runner path.
$75k–$100k path
Grant plan targets 30–50 fixtures, 10–12 failure classes, external validation notes, Docker/API hardening, and a technical report.
Ask ladder
The $75k–$100k range becomes credible when tied to external validation, expanded benchmarks, and a technical report.
From docs to validation
The next movement is not more words. It is reproducible evidence: more fixtures, more reports, more validators.
4-month execution plan
A clear delivery path for moving from early artifact to externally reproducible benchmark-backed research infrastructure.
Taxonomy + fixture design
- 10–12 failure classes
- fixture metadata
- initial 12–15 cases
Benchmark expansion
- 25–35 fixtures
- positive/negative controls
- JSON + Markdown reports
External validation
- 2+ validation notes
- Docker demo hardening
- reproduction issues tracked
Technical report
- 30–50 fixtures
- final benchmark report
- publication-ready draft
The next proof is not louder claims. It is independent reproduction.
The strongest $75k–$100k case: 30–50 fixtures, 2–5 external validation notes, machine-readable expected findings, and a technical report that states both results and limits.
We request $75,000–$100,000 to expand CML into a benchmark-backed, externally reproducible research artifact for causal-validity checking in agentic AI workflows.