Harden parsers and share pack validation

Author: josediegorobles

CHANGELOG.md

@@ -6,11 +6,13 @@
 
 - Added a real `ai` feature implementation for `summarize --input report.json --provider openai` using an OpenAI-compatible `/chat/completions` endpoint configured by `BTC_RISK_LAB_AI_BASE_URL`, `BTC_RISK_LAB_AI_API_KEY`, and optional `BTC_RISK_LAB_AI_MODEL`.
 - Added a global `--offline` flag that turns network-backed commands into explanatory errors before any HTTP client is used.
+- Added property tests covering transaction, PSBT, and script analysis against arbitrary bytes and mutated valid fixtures.
 
 ### Changed
 
 - AI summary output is explicitly marked `AI-assisted draft` and the command sends only the already-produced btc-risk-lab JSON report to the configured provider.
 - Builds compiled without `--features ai` now fail the summary command with `compiled without ai feature`.
+- Extracted shared pack input validation, file presence checks, file reading, and metadata helpers into `pack_common`.
 
 ### Security
 

Cargo.lock

@@ -31,4 +31,5 @@ tokio = { version = "1", default-features = false, features = ["rt-multi-thread"
 [dev-dependencies]
 assert_cmd = "2"
 predicates = "3"
+proptest = "1"
 tempfile = "3"

Cargo.toml

@@ -1,4 +1,5 @@
 pub mod analyzer;
+mod pack_common;
 pub mod policy_pack;
 pub mod report;
 pub mod review_pack;

src/lib.rs

@@ -0,0 +1,40 @@
+use std::{
+    fs,
+    path::{Path, PathBuf},
+};
+
+use anyhow::{bail, Context, Result};
+
+pub fn validate_input_dir(input_dir: &Path, pack_label: &str) -> Result<()> {
+    if !input_dir.is_dir() {
+        bail!(
+            "{pack_label} input must be a directory: {}",
+            input_dir.display()
+        );
+    }
+
+    Ok(())
+}
+
+pub fn optional_file(input_dir: &Path, file: &str) -> Option<PathBuf> {
+    let path = input_dir.join(file);
+    path.exists().then_some(path)
+}
+
+pub fn has_file(input_dir: &Path, file: &str) -> bool {
+    input_dir.join(file).exists()
+}
+
+pub fn read_to_string(path: &Path, context: impl FnOnce() -> String) -> Result<String> {
+    fs::read_to_string(path).with_context(context)
+}
+
+pub fn file_name(path: &Path) -> String {
+    path.file_name()
+        .map(|name| PathBuf::from(name).display().to_string())
+        .unwrap_or_else(|| path.display().to_string())
+}
+
+pub fn is_metadata_artifact(artifact: &str) -> bool {
+    artifact == "metadata"
+}

src/pack_common.rs

@@ -1,14 +1,12 @@
-use std::{
-    fs,
-    path::{Path, PathBuf},
-};
+use std::path::Path;
 
-use anyhow::{bail, Context, Result};
+use anyhow::Result;
 use serde::{Deserialize, Serialize};
 use serde_json::Value;
 
 use crate::{
     analyzer::{RiskLevel, RiskWarning, SummaryItem},
+    pack_common,
     review_pack::{
         analyze_review_pack, ArtifactSummary, CrossArtifactFinding, DetectedArtifact,
         ReviewPackReport,
@@ -49,12 +47,7 @@ pub struct PolicyFinding {
 }
 
 pub fn analyze_policy_pack(input_dir: &Path) -> Result<PolicyPackReport> {
-    if !input_dir.is_dir() {
-        bail!(
-            "policy pack input must be a directory: {}",
-            input_dir.display()
-        );
-    }
+    pack_common::validate_input_dir(input_dir, "policy pack")?;
 
     let review_pack = analyze_review_pack(input_dir)?;
     let evidence_documents = read_evidence_documents(input_dir)?;
@@ -119,16 +112,16 @@ fn read_evidence_documents(input_dir: &Path) -> Result<Vec<EvidenceDocument>> {
     let mut documents = Vec::new();
 
     for spec in evidence_specs() {
-        let path = input_dir.join(spec.file);
-        if !path.exists() {
+        let Some(path) = pack_common::optional_file(input_dir, spec.file) else {
             continue;
-        }
+        };
 
-        let input = fs::read_to_string(&path)
-            .with_context(|| format!("failed to read evidence document {}", path.display()))?;
+        let input = pack_common::read_to_string(&path, || {
+            format!("failed to read evidence document {}", path.display())
+        })?;
         documents.push(EvidenceDocument {
             artifact: spec.artifact.to_owned(),
-            file: file_name(&path),
+            file: pack_common::file_name(&path),
             format: spec.format.to_owned(),
             summary: summarize_evidence(spec.format, &input),
         });
@@ -424,7 +417,7 @@ fn has_policy_notes(documents: &[EvidenceDocument]) -> bool {
 fn has_metadata(documents: &[EvidenceDocument]) -> bool {
     documents
         .iter()
-        .any(|document| document.artifact == "metadata")
+        .any(|document| pack_common::is_metadata_artifact(&document.artifact))
 }
 
 fn push_detected_once(artifacts: &mut Vec<DetectedArtifact>, artifact: DetectedArtifact) {
@@ -436,12 +429,6 @@ fn push_detected_once(artifacts: &mut Vec<DetectedArtifact>, artifact: DetectedA
     }
 }
 
-fn file_name(path: &Path) -> String {
-    path.file_name()
-        .map(|name| PathBuf::from(name).display().to_string())
-        .unwrap_or_else(|| path.display().to_string())
-}
-
 fn artifact_label(input: &str) -> String {
     if input == "psbt" {
         return "PSBT".to_owned();