chore(p0): retire regress, the skill router, the module audit tool and vocabulary.sc

Off every active line (PATHFINDER-2026-10-13/03-unified-proposal.md, cuts).
Deleted: src/openultrasast/regress/ and the deep-mode REGRESS branch (deep
now records a regress_retired degradation and equals standard; proof is the
search-with-proof arm's job), the sandbox docker runner and probe that only
regress used, --fail-on worth-fixing, RegressConfig, the two ledger helpers
only regress called; skills.py and index.py with the hunter pool's skill
routing; model/audit.py; cpg/queries/vocabulary.sc with
benchmarks/push/ontology_gap.py. Docs and the scan skill describe the
retirement. Net: 3,895 lines deleted, 86 added. Targeted tests: 180 passed,
28 skipped; 3,331 tests collect.

Co-Authored-By: Claude Fable 5.1 <[email protected]>
This commit is contained in:
Marius Ciepluch
2026-10-13 10:50:01 +02:00
co-authored by Claude Fable 5.1
parent 94e96c6885
commit f5a904b7f2
50 changed files with 86 additions and 3895 deletions
+3 -13
View File
@@ -16,7 +16,7 @@ The CLI is thin. `HarnessRuntime` owns the lifecycle. It writes an event trace f
- `quick` runs STATIC;
- `standard` adds MAP;
- `deep` adds REGRESS.
- `deep` requests the REGRESS stage, which was retired on 2026-10-13: the scan records a `regress_retired` degradation and otherwise equals `standard`. Execution proof now comes from the search-with-proof arm.
```
policy_load ─▶ ruleset_load ─▶ policy_check (fail loud on an ungoverned CWE)
@@ -31,7 +31,7 @@ policy_load ─▶ ruleset_load ─▶ policy_check (fail loud on an
│ obligations absence bugs: an obligated operation with no governing guard
│ model one Joern CPG per repository; taint, dominance and configuration arbiters
│ tool hunter only with `[models] hunter`
─▶ REGRESS (deep) sandboxed regression snippets ─▶ verdicts.json, worth-fixing
─▶ REGRESS (deep) retired 2026-10-13, recorded as `regress_retired`
─▶ report.md, report.sarif, manifest.json
```
@@ -67,17 +67,7 @@ nothing, the LLM sees the enumerator's candidates (at most `MAX_JUDGED_CANDIDATE
Anything it affirms is reported at `suspicion`. Without a provider key, neither question is
asked. Only the graph's entailed findings are reported then.
**REGRESS** (`regress/`, `sandbox/`) runs only when `docker info` succeeds. Each promoted
candidate gets a small language recipe (Python, JavaScript, C). The recipe loads the
candidate's source inside a container with:
- no network;
- a read-only `/workspace`;
- a non-root user;
- the `[sandbox]` limits.
Before anything runs, a structural deny-list rejects snippets that mention sockets, curl, host
networking or writes under `/workspace`.
**REGRESS** was retired on 2026-10-13. Deep mode records a `regress_retired` degradation and otherwise equals standard. Execution proof is the search-with-proof arm's job (`search/verify.py`, bubblewrap sandbox, differential runs against owned canaries).
## Evidence ladder: how a false positive is eliminated