chore(release): sync develop to main for v0.14.0

.github/workflows/cla-check.yml

@@ -12,6 +12,7 @@ permissions:
 jobs:
   cla-check:
     runs-on: ubuntu-latest
+    if: "!contains(fromJSON('[\"app/dependabot\", \"app/renovate\", \"github-actions[bot]\"]'), github.event.pull_request.user.login)"
     steps:
       - name: Fetch & check CLA signature
         id: check

AGENT.md

@@ -19,7 +19,7 @@ semantic search engine (Brain Mode).
 ```bash
 cargo build              # Build (debug)
 cargo build --release    # Build (release)
-cargo test               # Run all 708 tests (default) / 714 (tree-sitter)
+cargo test               # Run all 934 tests (default) / 934 (tree-sitter)
 cargo clippy --all-targets -- -D warnings  # Lint (strict -D warnings)
 cargo fmt --all -- --check  # Format check
 ```
@@ -74,6 +74,7 @@ src/
 │   ├── llm.rs           # LLM API interaction
 │   ├── types.rs         # Severity, finding, and result types
 │   ├── diff_parser.rs   # Diff → FileChunk parsing
+│   ├── enclosing.rs     # Enclosing-scope control-flow context for review prompts
 │   ├── chunker.rs       # Auto-chunking large diffs
 │   ├── profiles.rs      # Quality profiles (strict/balanced/lax)
 │   ├── quality_gate.rs  # Quality gate thresholds + pass/fail
@@ -125,7 +126,7 @@ src/
 ## Testing
 
 ```bash
-cargo test               # 708 tests (default) / 714 (tree-sitter)
+cargo test               # 934 tests (default) / 934 (tree-sitter)
                          #   637 unit tests
                          #    16 CLI integration tests
                          #     6 config tests

CHANGELOG.md

@@ -7,8 +7,26 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.14.0] - 2026-08-28
+
+### Fixed
+
+- **Empty LLM responses from reasoning models.** Models like GLM can spend the entire `max_tokens` budget on chain-of-thought and return `content: ""` with `finish_reason: "length"`, which previously surfaced as a misleading `EOF while parsing` error. Cora now reads `finish_reason`/`reasoning_content`, automatically retries with a doubled budget (up to 32768), salvages JSON from reasoning text as a last resort, and reports an explicit "EMPTY response" error when nothing is recoverable (#536).
+
+- **Dead-code false positives from cross-crate method calls.** Method calls inside Rust `impl` blocks were never walked for call edges, and call targets stored raw AST text (`self.export_full`) that could not join against symbol names — 557 false positives on a 5-crate workspace (#519).
+- **Index root mismatch between CLI and MCP.** Running `cora index` inside a workspace member crate created a separate project row from the one MCP resolved, so `index_status` reported 0 symbols despite a populated DB. Root resolution now prefers a `[workspace]` Cargo.toml and never climbs past a `.git` boundary (#522).
+- **`ignore.files` was not honored by the index.** Skip patterns only invalidated fingerprints; matched files were still indexed and surfaced in dead-code/review findings. They are now excluded from indexing entirely (#521).
+
+### Changed
+
+- **Default `max_tokens` raised from 4096 to 8192** to give reasoning models headroom above their chain-of-thought (#536).
+
 ### Changed
 
+- **Default `max_tokens` raised from 4096 to 8192** to give reasoning models headroom above their chain-of-thought (#536).
+- **Dead-code now skips public API surface by default** (`pub`/`export` items) — new `--include-pub` flag and MCP `include_pub_api` parameter opt back in (#520).
+- **Review prompts include enclosing control-flow scope.** Hunks touching branching constructs get the enclosing function from the post-image (120-line cap), plus an always-on guardrail against unverified reachability claims (#523).
+- **Incremental re-index reports honestly.** No-op runs print "Index up to date" with stored totals instead of "Indexed 0 symbols"; MCP `index_status` carries a root-mismatch hint (#522).
 - **Relicensed from MIT to Apache-2.0.** All 18 CodeCoraDev repositories now
   standardize on Apache-2.0 for patent grant protection and open-core model
   compatibility. Added CLA (Individual + Corporate) for contributor copyright

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cora-code"
-version = "0.13.0"
+version = "0.14.0"
 edition = "2024"
 description = "CLI-first AI code review — BYOK, diff/scan/branch, pre-commit hooks"
 license = "Apache-2.0"

SECURITY.md

@@ -45,6 +45,31 @@ Security-related areas of the codebase:
 - `src/hook/` — Pre-commit hook integration
 - `src/index/` — File access and SQLite storage
 
+## Threat Model: Adversarial Source-Code Comments (ALIBI)
+
+LLM-based reviewers are vulnerable to adversarial comments in the code under
+review that steer reviewer reasoning without changing program behavior —
+attack success exceeds 90% across 125 real-world vulnerabilities, with
+fabricated tool-result claims ("sanitizer passed", "already validated") being
+the most effective vector (arXiv:2607.24964).
+
+**Prompt-level defenses (telling the model to ignore comments) are proven
+ineffective against adaptive attacks.** Cora therefore uses architectural
+defenses:
+
+- **Claim flagging (always on)** — added comments asserting verification or
+  tool results are detected heuristically and injected into review context as
+  *untrusted claims*, never as facts.
+- **Comment sanitization (opt-in)** — set `review.sanitize-comments: true` in
+  `.cora.yaml` to strip comment bodies from added diff lines before the LLM
+  sees them. Line structure is preserved (`[comment removed]` markers), so
+  findings still map to real line numbers. Deterministic scanners (rules,
+  secrets, security patterns) always run on the *unsanitized* diff.
+- Sanitization is heuristic (line-comment markers `//`, leading `#`, `--`,
+  `;`); block comments (`/* */`, `"""..."""`) are not currently stripped.
+
+Relevant code: `src/engine/comment_sanitizer.rs`.
+
 ## Responsible Disclosure
 
 We follow responsible disclosure principles:

docs/cli-reference.md

@@ -119,7 +119,8 @@ See [Code Intelligence](./code-intelligence) for detailed usage.
 | `cora affected` `<files...>` | Find test files affected by source changes |
 | `cora affected --stdin` | Read changed files from stdin (pipe from `git diff --name-only`) |
 | `cora affected --filter` `"*test*"` | Custom test file glob pattern |
-| `cora dead-code` | Detect dead code — functions/methods with zero callers |
+| `cora dead-code` | Detect dead code — functions/methods with zero callers (public API surface skipped by default; honors ignore.files via the index) |
+| `cora dead-code --include-pub` | Include public API surface (pub/export items) in results |
 | `cora dead-code --include-tests` | Include test functions in results |
 | `cora dead-code --min-lines N` | Filter out tiny functions |
 | `cora query` `"main -> *"` | Query the code graph with simple patterns |

docs/configuration.md

@@ -57,6 +57,9 @@ review:
   system_prompt: "You are a senior code reviewer."
   # system_prompt_file: ./review-prompt.md
   response_format: json_object
+  # Strip comments from added diff lines before the LLM sees them
+  # (ALIBI defense, arXiv:2607.24964). Claim flagging is always on.
+  sanitize_comments: false
   static_analysis:
     auto_clippy: false       # auto-run `cargo clippy` (Rust only)
     clippy_output_file: ""   # or read clippy output from file

src/commands/scan.rs

@@ -117,11 +117,15 @@ pub async fn execute_scan(
     // 2b. Run index-powered deterministic scans (unused imports, dead code)
     // These work even without LLM and add findings to the final report.
     let root_abs = root.canonicalize().unwrap_or_else(|_| root.clone());
+    // Same merged exclusion set as `cora index` (#521).
+    let mut index_skip = config.ignore.files.clone();
+    index_skip.extend(config.rules_config.index_skip_files.iter().cloned());
+    index_skip.dedup();
     let index_findings = crate::engine::index_scanner::scan_project_index(
         &root_abs,
         &files,
         config.rules_config.max_findings,
-        &config.rules_config.index_skip_files,
+        &index_skip,
     );
     if !index_findings.is_empty() {
         eprintln!(

src/commands/watch.rs

@@ -37,9 +37,13 @@ pub fn run_watch(
     // Load skip patterns + brain embedding backend from config
     let config =
         crate::config::loader::load_config(config_path, None, None, None, None, false).ok();
-    let skip_patterns: Option<Vec<String>> = config
-        .as_ref()
-        .map(|c| c.rules_config.index_skip_files.clone());
+    // Same merged exclusion set as `cora index` (#521).
+    let skip_patterns: Option<Vec<String>> = config.as_ref().map(|c| {
+        let mut pats = c.ignore.files.clone();
+        pats.extend(c.rules_config.index_skip_files.iter().cloned());
+        pats.dedup();
+        pats
+    });
 
     // Resolve embedding backend
     let brain_mode = config

src/config/schema.rs

@@ -43,6 +43,9 @@ pub struct Config {
     pub cache_ttl: u64,
     /// Static analysis context injection for reviews.
     pub static_analysis: StaticAnalysisConfig,
+    /// Strip comments from added diff lines before the LLM sees them
+    /// (ALIBI defense, arXiv:2607.24964).
+    pub sanitize_comments: bool,
     /// Rule engine configuration.
     pub rules_config: RulesConfig,
     /// Context chain configuration — cross-file dependency extraction.
@@ -143,10 +146,11 @@ impl Default for Config {
             response_format: "none".to_string(),
             review_system_prompt_override: None,
             review_system_prompt_file: None,
+            sanitize_comments: false,
             scan_system_prompt_override: None,
             scan_system_prompt_file: None,
             temperature: 0.0,
-            max_tokens: 4096,
+            max_tokens: 8192, // #536: reasoning models need headroom above chain-of-thought
             max_tokens_param: "auto".to_string(),
             timeout: 600,
             cache_ttl: 1440, // 24h in minutes
@@ -404,6 +408,10 @@ pub struct ReviewSection {
     /// Static analysis context injection (e.g., clippy output).
     #[serde(skip_serializing_if = "Option::is_none")]
     pub static_analysis: Option<StaticAnalysisConfig>,
+    /// Strip comments from added diff lines before the LLM sees them
+    /// (ALIBI defense, arXiv:2607.24964).
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub sanitize_comments: Option<bool>,
     /// Context chain configuration (cross-file dependency extraction).
     #[serde(skip_serializing_if = "Option::is_none")]
     pub context_chain: Option<crate::engine::context::types::ContextConfig>,
@@ -658,6 +666,9 @@ impl CoraFile {
             if let Some(sa) = &r.static_analysis {
                 config.static_analysis.clone_from(sa);
             }
+            if let Some(v) = r.sanitize_comments {
+                config.sanitize_comments = v;
+            }
             if let Some(cc) = &r.context_chain {
                 config.context_chain.clone_from(cc);
             }
@@ -1211,6 +1222,7 @@ review:
                 system_prompt: None,
                 system_prompt_file: None,
                 static_analysis: None,
+                sanitize_comments: None,
                 context_chain: None,
             }),
             ..Default::default()
@@ -1228,6 +1240,7 @@ review:
                 system_prompt: Some("Custom prompt here.".to_string()),
                 system_prompt_file: None,
                 static_analysis: None,
+                sanitize_comments: None,
                 context_chain: None,
             }),
             ..Default::default()
@@ -1248,6 +1261,7 @@ review:
                 system_prompt: None,
                 system_prompt_file: Some("prompts/review.md".to_string()),
                 static_analysis: None,
+                sanitize_comments: None,
                 context_chain: None,
             }),
             ..Default::default()
@@ -1326,7 +1340,7 @@ scan:
     #[test]
     fn config_default_max_tokens() {
         let cfg = Config::default();
-        assert_eq!(cfg.max_tokens, 4096);
+        assert_eq!(cfg.max_tokens, 8192);
     }
 
     #[test]
@@ -1388,7 +1402,7 @@ llm:
         cora.merge_into(&mut cfg).unwrap();
         assert_eq!(cfg.temperature, 0.7);
         // Other LLM fields should remain at defaults
-        assert_eq!(cfg.max_tokens, 4096);
+        assert_eq!(cfg.max_tokens, 8192);
         assert_eq!(cfg.timeout, 600);
         assert_eq!(cfg.cache_ttl, 1440);
     }