marius/Open_Ultra_SAST
1
0
Fork 0
mirror of https://github.com/norandom/OpenUltraSAST.git synced 2026-10-16 03:49:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(learn): candidate feature schema and builders (decision-engine task 1)

Browse Source 
learn/schema.py is the closed, typed allow-list of candidate features: 66 features over
11 instruments (language one-hot, quick rules with 23 mechanism buckets and a prior count,
engine, repo-facts, source span, entry-point distance, roles, delta, verify, agree, model
sinks), two profiles (static 57; plane adds verify, agree and model sinks for 66), schema_version and a feature-set digest.
No repository, path, file, function, rule or commit identity is a feature; site_match,
in_fix_range, fixed-pin alerts, case.json fields and costs are refused. Missing is not zero:
each record carries per-instrument state (ran | none | failed | not_applicable), and every
feature of an instrument that did not run is null.

learn/features.py builds records on the host (build_for_scan, whole repository or base..head
delta); plane/tasks/features.py builds them per case in a Run from the delivered artifacts,
dropping site_match and in_fix_range on read and ignoring fixed-pin alerts. The generator
adds a model-free `features` entry per case before `remember`, which stores the records as
`features` rows (re-validated). memory KINDS gains features, label, decision, experiment,
arm_outcome and experiment_result. The committed validation-46 manifests are regenerated
(entries appended; the first Run's entries are unchanged). The runner image must be rebuilt
and re-pinned before a Run executes the new task.

Proof on real artifacts (scratch store, not committed): the recorded validation-46 run gives
75 feature records over 15 cases (43 verify candidates plus vulnerable-pin alerts), each
case's vulnerable pin exported and read (0.5k-9k files); remember stored 75 features rows
beside 43 verdict, 84 alert, 61 unit_cost, 15 facts and 8 coverage rows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EmM1j46qgKMh3zqD1QMwEo
This commit is contained in:
Marius Ciepluch
2026-09-30 18:38:04 +02:00
co-authored by Claude Opus 5.5
parent edcf145b00
commit d1896cb8e7
17 changed files with 2129 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files
benchmarks/measurements/2026-09-08-module-audit.json
+54 -4
View File
@@ -1,7 +1,7 @@
{
"modules": 135,
"modules": 139,
"counts": {
"load_bearing": 128,
"load_bearing": 132,
"standalone_capability": 7
},
"orphaned": [],
@@ -346,6 +346,33 @@
"skills"
]
},
{
"module": "learn",
"classification": "load_bearing",
"reason": "reachable from gate, map_gate, pair_gate or the CLI",
"importers": [
"learn.features",
"plane.tasks.features",
"plane.tasks.remember"
]
},
{
"module": "learn.features",
"classification": "load_bearing",
"reason": "imported by plane.tasks.features",
"importers": [
"plane.tasks.features"
]
},
{
"module": "learn.schema",
"classification": "load_bearing",
"reason": "reachable from gate, map_gate, pair_gate or the CLI",
"importers": [
"learn.features",
"plane.tasks.remember"
]
},
{
"module": "map_gate",
"classification": "load_bearing",
@@ -361,6 +388,7 @@
"gate",
"hunter_tools",
"improve.evolve",
"learn.features",
"map_gate",
"model.scan",
"pairs",
@@ -386,6 +414,7 @@
"cli",
"cpg.backend",
"cpg.session",
"learn.features",
"model.candidates",
"model.config_value",
"model.dominance",
@@ -401,6 +430,7 @@
"model.taint",
"plane.budget",
"plane.generate",
"plane.tasks.features",
"plane.tasks.repo_facts",
"preprocess",
"push.assistance",
@@ -626,10 +656,12 @@
"reason": "reachable from gate, map_gate, pair_gate or the CLI",
"importers": [
"cli",
"learn.features",
"model.candidates",
"model.report",
"model.scan",
"model.specs"
"model.specs",
"plane.tasks.features"
]
},
{
@@ -768,11 +800,13 @@
"reason": "ai-service-plane Req 4.1: the task entrypoint package the runner imports by name; runner integration pending (task 3)",
"importers": [
"cli",
"learn.features",
"plane.engine_alerts",
"plane.generate",
"plane.memory",
"plane.tasks.agree",
"plane.tasks.alerts",
"plane.tasks.features",
"plane.tasks.remember"
]
},
@@ -787,9 +821,17 @@
"classification": "load_bearing",
"reason": "harnessx-removal Req 6.3: quick-mode alerts on a case's vulnerable and fixed pins, model-free",
"importers": [
"plane.engine_alerts"
"learn.features",
"plane.engine_alerts",
"plane.tasks.features"
]
},
{
"module": "plane.tasks.features",
"classification": "load_bearing",
"reason": "learned-decision-engine Req 1.1-1.3/7.2: one allow-listed feature record per candidate, model-free",
"importers": []
},
{
"module": "plane.tasks.loop",
"classification": "load_bearing",
@@ -811,8 +853,10 @@
"classification": "load_bearing",
"reason": "ai-service-plane Req 5.1/5.2: source-only functions and callers per file; runner integration pending",
"importers": [
"learn.features",
"plane.generate",
"plane.tasks.alerts",
"plane.tasks.features",
"plane.tasks.remember"
]
},
@@ -872,6 +916,7 @@
"hunter_tools",
"improve.evolve",
"index",
"learn.features",
"map_gate",
"mapping",
"model.candidates",
@@ -879,6 +924,7 @@
"model.scan",
"pairs",
"plane.tasks.alerts",
"plane.tasks.features",
"plane.tasks.repo_facts",
"push.runner",
"rank",
@@ -1148,8 +1194,10 @@
"improve.evolve",
"improve.memory",
"improve.validator",
"learn.features",
"pairs",
"plane.tasks.alerts",
"plane.tasks.features",
"plane.tasks.loop",
"stage_processors"
]
@@ -1229,6 +1277,7 @@
"cli",
"cpg.capability",
"hunter_tools",
"learn.features",
"mapping",
"model.candidates",
"model.layout",
@@ -1283,6 +1332,7 @@
"importers": [
"cli",
"hunter_tools",
"learn.features",
"mapping",
"model.candidates",
"model.layout",