cschanhniem · Jul 9, 2026 · Jul 9, 2026
diff --git a/Makefile b/Makefile
@@ -1,4 +1,4 @@
-.PHONY: help demo test lint ci clean bench-leakage bench-multi-negatives bench-baseline bench-hidden-active bench-cluster-split bench-expert-ablation bench-expert-ablation-500 bench-selectivity bench-feature-decomp bench-gate bench-easy-baseline bench-charge-matched bench-order-dependent bench-precision-at-k bench-per-family bench-simulation-ablation bench-simulation-ablation-within-amp simulation-gate regenerate-all generate phase3 pilot validate-scoring validate-scoring-phase3 validate-scoring-strict external-pilot pilot-confident presynth-qc gold-standard diversity synthesis-order novelty-broad external-consensus questionnaire gate-check ip-report benchmark-card wave0-5-gate-check wave0-5-novelty-audit wave0-5-novelty-audit-v2 wave0-5-panel wave0-5-evidence wave0-5-fill-external wave0-5b-generate wave0-5b-filter recalibration-engine recalibration-engine-dry-run validate-policy-version generate-synthetic-lab-results bump-policy-version generate-review-packet failed-candidate-report
+.PHONY: help demo test lint ci clean bench-leakage bench-multi-negatives bench-baseline bench-hidden-active bench-cluster-split bench-expert-ablation bench-expert-ablation-500 bench-selectivity bench-feature-decomp bench-gate bench-easy-baseline bench-charge-matched bench-order-dependent bench-precision-at-k bench-per-family bench-simulation-ablation bench-simulation-ablation-within-amp simulation-gate regenerate-all generate phase3 pilot validate-scoring validate-scoring-phase3 validate-scoring-strict external-pilot pilot-confident presynth-qc gold-standard diversity synthesis-order novelty-broad external-consensus questionnaire gate-check ip-report benchmark-card wave0-5-gate-check wave0-5-novelty-audit wave0-5-novelty-audit-v2 wave0-5-panel wave0-5-evidence wave0-5-fill-external wave0-5b-generate wave0-5b-filter recalibration-engine recalibration-engine-dry-run validate-policy-version generate-synthetic-lab-results bump-policy-version generate-review-packet failed-candidate-report safe-publication-filter
 
 PYTHON := $(shell [ -f .venv/bin/python ] && echo .venv/bin/python || echo python3)
 PYTEST  := $(shell [ -f .venv/bin/pytest ] && echo .venv/bin/pytest || echo pytest)
@@ -602,3 +602,9 @@ failed-candidate-report:
 		--out-json outputs/failed_candidate_report.json \
 		--out-md outputs/failed_candidate_report.md \
 		--validate-rejection-codes
+
+safe-publication-filter:
+	PYTHONPATH=src $(PYTHON) scripts/safe_publication_filter.py \
+		--input examples/safe_publication_filter_example_input.json \
+		--out-json outputs/safe_publication_filter_result.json \
+		--out-md outputs/safe_publication_filter_result.md
diff --git a/docs/evidence/METRICS_CURRENT.md b/docs/evidence/METRICS_CURRENT.md
@@ -5,7 +5,9 @@ Machine-readable snapshot: `outputs/metrics_snapshot.json` regenerated with `mak
 > **Purpose:** One authoritative table of current pipeline metrics. If any doc disagrees
 > with this file, this file wins. Updated whenever benchmark/benchmark config changes.
 >
-> **Last updated:** 2026-07-09 (Phase F report generator — v0.5.75)
+> **Last updated:** 2026-07-09 (Phase F safe-publication filter — v0.5.76)
+> **New in v0.5.76:** Safe-publication filter (F5) — CLI that reads candidate panel data with safety metadata and checks each candidate against publication safety constraints before external release. FAIL-CLOSED: defaults to reject unless all checks pass. Checks: dry_lab_only (must be true), proof_ladder_level (≤4), toxicity_screened, hemolysis_screened, dual_use_reviewed (all must be true). Produces structured JSON filter result and human-readable Markdown summary. 33 tests. 2758 total.
+>
 > **New in v0.5.75:** Failed-candidate report generator (F4) — CLI that reads candidate rejection data + taxonomy, produces structured JSON report with summary aggregation (by category, severity, stage, evidence impact) and human-readable Markdown summary. 26 tests. 2724 total.
 >
 > **New in v0.5.74:** Rejection reason taxonomy schema (F3) — machine-readable taxonomy with severity, evidence_impact, and applies_at_stage fields. Comprehensive REJECTION_TAXONOMY.md doc. 19 tests. 2698 total.
@@ -60,7 +62,7 @@ Machine-readable snapshot: `outputs/metrics_snapshot.json` regenerated with `mak
 > **New in v0.5.31:** Added dipeptide-order features for sequence-order awareness. `dipeptide_order_score` achieves AUROC 0.7861 on AMP-vs-scrambled discrimination — the strongest order-dependent feature in the pipeline. Only 7/31 features survive scrambling (amphipathicity/helix-wheel + dipeptide). All composition features are purely position-independent (exactly 0.5000 AUROC on scrambled test).
 > **New in v0.5.30:** Easy baseline benchmark added — charge density alone (AUROC 0.8166) outperforms the full pipeline ensemble (0.7792) on AMP-vs-Swiss-Prot-decoy discrimination. Honest finding documented: expected because pipeline optimizes for safety, not raw discrimination.
 > **New in v0.5.29:** Expanded benchmark to 500 AMPs + 500 composition-matched decoys (n=1000). AUROC 0.7792 (CI₉₅: 0.7505–0.8065) confirms signal generalizes. Cluster-aware CI: 0.746–0.8102. Representative AUROC: 0.778. Standard benchmark (n=191) retained for backward comparison.
-> **Pipeline version:** v0.5.75
+> **Pipeline version:** v0.5.76
 > **Branch:** main
 
 ---

diff --git a/docs/research/ROADMAP.md b/docs/research/ROADMAP.md
@@ -1,5 +1,49 @@
 # Roadmap
 
+## v0.5.76 — Loop 75: Phase F F5 — Safe-Publication Filter ✓ (2026-07-09)
+
+Safe-publication filter: CLI that reads a candidate panel with safety metadata
+and checks each candidate against publication safety constraints before external
+release. FAIL-CLOSED: defaults to reject unless all checks pass.
+
+Changes:
+- `scripts/safe_publication_filter.py` (F5) — Standalone CLI that loads candidate
+  panel data with publication safety metadata and applies global-level checks
+  (dry_lab_only must be true, proof_ladder_level ≤ 4, toxicity_screened,
+  hemolysis_screened, dual_use_reviewed all true) plus per-candidate checks
+  (proof_ladder_level ≤ 4). Produces structured JSON with filter metadata,
+  global checks, per-candidate results, and summary. Exit code 0 if all
+  checks pass, 1 if any fail (FAIL-CLOSED), 2 for input errors.
+- `schemas/safe_publication_filter.schema.json` (F5) — JSON Schema Draft 2020-12
+  for the filter result output. Validates filter_metadata, global_checks
+  (3 required sub-check groups), per-candidate results, summary counts, and
+  caveats.
+- `examples/safe_publication_filter_example_input.json` (F5) — Toy input
+  with 4 candidates spanning PASS (pll=2, pll=1), WARN (pll=4 at boundary),
+  and FAIL (pll=5 too high). Clearly marked EXAMPLE.
+- `Makefile` — Added `safe-publication-filter` target.
+- `tests/evidence/test_safe_publication_filter.py` — 33 tests covering:
+  global check success/failure paths (dry_lab_only, proof_ladder_level,
+  all safety checks individually), per-candidate checks (valid, missing,
+  too-high, boundary/warn), mixed candidate results, filter result summary
+  counts, overall PASS/FAIL logic, markdown output sections, input loading
+  (valid/missing/invalid JSON), CLI exit codes (0, 1, 2), CLI output
+  (JSON + Markdown), caveat generation, metadata structure, and example
+  input round-trip validation.
+
+Honest boundaries:
+- The filter checks structural publication safety constraints, not scientific
+  validity. A PASS result does not mean the candidate is biologically active
+  or safe.
+- dry_lab_only attestation is a safety-constraint field; it relies on the
+  generator's honesty.
+- Proof ladder levels are self-reported by the pipeline; independent
+  verification is recommended.
+- Safety checks (toxicity, hemolysis, dual-use) are computational assessments
+  and require qualified human review.
+- A FAIL result does not mean the candidate is biologically inactive — it
+  means publication safety checks are not satisfied.
+
 ## v0.5.75 — Loop 74: Phase F F4 — Failed-Candidate Report Generator ✓ (2026-07-09)
 
 Failed-candidate report generator: CLI that reads candidate rejection data validated

diff --git a/examples/safe_publication_filter_example_input.json b/examples/safe_publication_filter_example_input.json
@@ -0,0 +1,39 @@
+{
+  "source": "wave0.5-example",
+  "pipeline_version": "v0.5.75",
+  "report_type": "safe_publication_input",
+  "description": "EXAMPLE safe-publication filter input. NOT a real candidate panel — illustrative only.",
+  "dry_lab_only": true,
+  "proof_ladder_level": 2,
+  "safety_checks": {
+    "toxicity_screened": true,
+    "hemolysis_screened": true,
+    "dual_use_reviewed": true
+  },
+  "candidates": [
+    {
+      "candidate_id": "EXAMPLE-CAND-001",
+      "sequence": "ACDEFGHIKLMNPQRSTVWY",
+      "ensemble_score": 0.85,
+      "proof_ladder_level": 2
+    },
+    {
+      "candidate_id": "EXAMPLE-CAND-002",
+      "sequence": "KLAKLAKKLAKLAK",
+      "ensemble_score": 0.80,
+      "proof_ladder_level": 4
+    },
+    {
+      "candidate_id": "EXAMPLE-CAND-003",
+      "sequence": "GIGKFLHSAKKWGKAFVGEIMNS",
+      "ensemble_score": 0.78,
+      "proof_ladder_level": 5
+    },
+    {
+      "candidate_id": "EXAMPLE-CAND-004",
+      "sequence": "RRWQWRMKKLG",
+      "ensemble_score": 0.72,
+      "proof_ladder_level": 1
+    }
+  ]
+}
diff --git a/schemas/safe_publication_filter.schema.json b/schemas/safe_publication_filter.schema.json
@@ -0,0 +1,163 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://openamp-foundry.org/schemas/safe_publication_filter.schema.json",
+  "title": "OpenAMP Safe Publication Filter Result",
+  "description": "JSON Schema for the safe-publication filter result. Validates the output of scripts/safe_publication_filter.py before any external release. FAIL-CLOSED: the filter defaults to reject unless all publication safety checks pass.",
+  "type": "object",
+  "required": [
+    "filter_metadata",
+    "global_checks",
+    "global_pass",
+    "candidates",
+    "summary",
+    "caveats"
+  ],
+  "properties": {
+    "filter_metadata": {
+      "type": "object",
+      "required": [
+        "generated_at",
+        "filter_type",
+        "filter_version",
+        "input_source",
+        "pipeline_version"
+      ],
+      "properties": {
+        "generated_at": {
+          "type": "string",
+          "format": "date-time",
+          "description": "ISO 8601 timestamp when the filter was run"
+        },
+        "filter_type": {
+          "type": "string",
+          "enum": ["safe_publication_filter"],
+          "description": "Filter type identifier"
+        },
+        "filter_version": {
+          "type": "string",
+          "pattern": "^\\d+\\.\\d+\\.\\d+$",
+          "description": "Filter schema version (semver)"
+        },
+        "input_source": {
+          "type": "string",
+          "description": "Source identifier of the input data"
+        },
+        "pipeline_version": {
+          "type": "string",
+          "description": "OpenAMP pipeline version"
+        }
+      }
+    },
+    "global_checks": {
+      "type": "object",
+      "description": "Global-level publication safety checks applied to the entire batch",
+      "required": ["dry_lab_only", "global_proof_ladder_level", "safety_checks"],
+      "properties": {
+        "dry_lab_only": {
+          "type": "object",
+          "required": ["required", "actual", "passed", "description"],
+          "properties": {
+            "required": { "type": "boolean", "enum": [true] },
+            "actual": { "type": "boolean" },
+            "passed": { "type": "boolean" },
+            "description": { "type": "string" },
+            "reason": { "type": "string" }
+          }
+        },
+        "global_proof_ladder_level": {
+          "type": "object",
+          "required": ["max_allowed", "actual", "passed", "description"],
+          "properties": {
+            "required": { "type": "string" },
+            "max_allowed": { "type": "integer", "minimum": 1, "maximum": 8 },
+            "actual": { "type": "integer", "minimum": 1, "maximum": 8 },
+            "passed": { "type": "boolean" },
+            "description": { "type": "string" },
+            "reason": { "type": "string" }
+          }
+        },
+        "safety_checks": {
+          "type": "object",
+          "required": ["all_passed", "checks"],
+          "properties": {
+            "all_passed": { "type": "boolean" },
+            "checks": {
+              "type": "object",
+              "minProperties": 3,
+              "additionalProperties": {
+                "type": "object",
+                "required": ["required", "actual", "passed"],
+                "properties": {
+                  "required": { "type": "boolean", "enum": [true] },
+                  "actual": { "type": "boolean" },
+                  "passed": { "type": "boolean" },
+                  "reason": { "type": "string" }
+                }
+              }
+            }
+          }
+        }
+      }
+    },
+    "global_pass": {
+      "type": "boolean",
+      "description": "Whether all global-level checks passed. If false, the filter FAILS regardless of per-candidate results."
+    },
+    "candidates": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "required": ["candidate_id", "overall", "checks"],
+        "properties": {
+          "candidate_id": { "type": "string" },
+          "overall": {
+            "type": "string",
+            "enum": ["PASS", "FAIL", "WARN"],
+            "description": "PASS: all checks pass, FAIL: any check fails, WARN: passes but at boundary"
+          },
+          "checks": {
+            "type": "object",
+            "minProperties": 1,
+            "additionalProperties": {
+              "type": "object",
+              "required": ["value", "max_allowed", "passed"],
+              "properties": {
+                "value": { "type": ["integer", "null"] },
+                "max_allowed": { "type": "integer" },
+                "passed": { "type": "boolean" },
+                "reason": { "type": "string" }
+              }
+            }
+          },
+          "issues": {
+            "type": "array",
+            "items": { "type": "string" }
+          }
+        }
+      }
+    },
+    "summary": {
+      "type": "object",
+      "required": ["total_candidates", "passed", "failed", "overall_filter_pass"],
+      "properties": {
+        "total_candidates": { "type": "integer", "minimum": 0 },
+        "passed": { "type": "integer", "minimum": 0 },
+        "failed": { "type": "integer", "minimum": 0 },
+        "warned": { "type": "integer", "minimum": 0 },
+        "overall_filter_pass": {
+          "type": "boolean",
+          "description": "True only if ALL global checks pass AND NO candidates FAIL. FAIL-CLOSED."
+        }
+      }
+    },
+    "caveats": {
+      "type": "array",
+      "minItems": 1,
+      "items": { "type": "string" }
+    },
+    "dry_lab_only": {
+      "type": "boolean",
+      "description": "Copied from input — all outputs are dry-lab computational hypotheses, not biological proof."
+    }
+  }
+}