mirror of https://github.com/safishamsi/graphify.git synced 2026-09-13 00:55:53 +00:00

Files

T

Christopher BeaulieuandClaude Opus 4.7 5685984110 evaluation(rsl-siege-manager): capture dry-run results + verdict

Adds the rsl-siege-manager evaluation under a new `evaluations/` convention:

- `evaluations/README.md` — directory index for future evaluations.
- `evaluations/rsl-siege-manager/runbook.md` — methodology used (relocated
  from siege-web's `docs/experiments/graphify-dry-run.md`).
- `evaluations/rsl-siege-manager/results.md` — decision log with quoted
  findings and acid-test scoring. Verdict: do not adopt.
- `evaluations/rsl-siege-manager/graphify-out/` and `…-with-tests/` — raw
  artifacts (GRAPH_REPORT.md, graph.html, graph.json, manifest.json) so
  quoted findings stay auditable. AST cache excluded.

Also extends `.gitignore` with a scoped exception so the rule that hides
fresh `graphify-out/` from a developer's working checkout still applies,
while committed evaluation artifacts under `evaluations/` are preserved.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

2026-05-15 11:08:27 -04:00

rsl-siege-manager

evaluation(rsl-siege-manager): capture dry-run results + verdict

2026-05-15 11:08:27 -04:00

README.md

evaluation(rsl-siege-manager): capture dry-run results + verdict

2026-05-15 11:08:27 -04:00

README.md

Graphify Evaluations

Per-codebase evaluations of graphify — runbook + captured artifacts + a written verdict.

Each subdirectory is one evaluation:

A runbook.md capturing the methodology used at the time.
A results.md documenting the run output, the acid-test scoring, and the adopt / don't-adopt decision.
The raw graphify-out/ directories (often two: one with tests excluded, one with tests included) so quoted findings remain auditable.

Evaluations

Codebase	Verdict	Date
`rsl-siege-manager/`	Do not adopt	2026-05-15

🤖 Generated by Claude Code on behalf of @cbeaulieu-gt