feat: add brick_callers and brick_uncalled caller-status queries #4

Merged

trevon merged 2 commits from feat/caller-status into feat/cpp-rust-resolution

2026-09-22 21:15:02 +00:00

trevon commented

2026-09-22 21:13:03 +00:00

Owner

Summary

Stacked on #3 (feat/cpp-rust-resolution); review and merge that first. This adds agent-facing MCP queries that answer "is this function used?" without ever turning an unresolved call graph into a dead-code verdict.

brick_callers {node_id, assume_closed_world?} returns one of three statuses:
- resolved, with the complete caller edges;
- unknown, with the reasons callers may exist;
- unreferenced.
It also returns an evidence basis (heuristic or compiler), the language's metrics.resolution coverage, and a fixed caveat.
brick_uncalled {language?, assume_closed_world?, include_unknown?, limit?} returns per-language status counts and a deterministic list of review candidates.
Reasons for unknown:
- same-name unresolved call sites, with counts by reason and example provenance (foreign-typed receivers and macros are excluded);
- references by name (callbacks, address-taken, macro bodies, attribute strings);
- virtual, trait-impl, external-base and outside-class members;
- framework attributes (#[test], #[tool], …);
- implicit invocation (constructors, operators, coroutine/range-for protocol);
- entry points and test code;
- parse-incomplete files;
- exported or external-linkage functions, unless assume_closed_world is set.
Also adds the research write-up (docs/research/cpp-rust-resolution.md): the audit, the literature found through paperbridge, the experiment contract, results, and limits.

Validation on YAMS (dead-code candidates checked with grep)

Open world: 2 candidates, both genuinely unused.
Closed world: 357 production candidates. A first random sample of 25 was 20 correct, and I fixed the rules behind the misses. An independent sample of 30 was then 29 correct, and I fixed that miss too.
On Fold, all 5 open-world candidates are unused.

Test plan

cargo test (103 tests), clippy and fmt pass.
tests/callers.rs covers each reason with positives and clean negatives.
The MCP CLI tests and the mcp_runner.py self-test/run pass with the two new tools.
./scripts/quality-gate.sh passes.

🤖 Generated with Claude Code

## Summary Stacked on #3 (`feat/cpp-rust-resolution`); review and merge that first. This adds agent-facing MCP queries that answer "is this function used?" without ever turning an unresolved call graph into a dead-code verdict. - **`brick_callers {node_id, assume_closed_world?}`** returns one of three statuses: - `resolved`, with the complete caller edges; - `unknown`, with the reasons callers may exist; - `unreferenced`. It also returns an evidence `basis` (`heuristic` or `compiler`), the language's `metrics.resolution` coverage, and a fixed caveat. - **`brick_uncalled {language?, assume_closed_world?, include_unknown?, limit?}`** returns per-language status counts and a deterministic list of review candidates. - **Reasons** for `unknown`: - same-name unresolved call sites, with counts by reason and example provenance (foreign-typed receivers and macros are excluded); - references by name (callbacks, address-taken, macro bodies, attribute strings); - virtual, trait-impl, external-base and outside-class members; - framework attributes (`#[test]`, `#[tool]`, …); - implicit invocation (constructors, operators, coroutine/range-for protocol); - entry points and test code; - parse-incomplete files; - exported or external-linkage functions, unless `assume_closed_world` is set. - Also adds the research write-up (`docs/research/cpp-rust-resolution.md`): the audit, the literature found through paperbridge, the experiment contract, results, and limits. ## Validation on YAMS (dead-code candidates checked with grep) - **Open world:** 2 candidates, both genuinely unused. - **Closed world:** 357 production candidates. A first random sample of 25 was 20 correct, and I fixed the rules behind the misses. An independent sample of 30 was then 29 correct, and I fixed that miss too. - On Fold, all 5 open-world candidates are unused. ## Test plan - [x] `cargo test` (103 tests), clippy and fmt pass. - [x] `tests/callers.rs` covers each reason with positives and clean negatives. - [x] The MCP CLI tests and the `mcp_runner.py` self-test/run pass with the two new tools. - [x] `./scripts/quality-gate.sh` passes. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

trevon added 2 commits

2026-09-22 21:13:03 +00:00

feat: add brick_callers and brick_uncalled caller-status queries d292b9f3f2

Agents asking "is this function used?" get resolved, unknown, or
unreferenced per function instead of a bare zero in-degree. Unknown lists
why callers may exist: same-name unresolved call sites (excluding foreign
receivers and macros), name references (callbacks, address-taken, macro
bodies), virtual/trait-impl/external-base/outside-class members, framework
attributes, implicit protocol invocations, entry points, test code,
incomplete parses, and, unless assume_closed_world is set, exported or
external-linkage functions. Unreferenced is a review candidate with an
explicit heuristic/compiler evidence basis, never a dead-code proof.

The queries are read-only over validated schema-4 models and return
complete canonical nodes and edges plus metrics.resolution coverage.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

docs: record C/C++ and Rust resolution research and validation

build / build-macos (pull_request) Has been cancelled

Details

build / release (pull_request) Has been cancelled

Details

build / build-linux (pull_request) Has been cancelled

Details

6d4335d3ae

Adds the audit, literature rationale, experiment contract, labeled-corpus
results, Fold/Paperbridge/Brick/YAMS coverage and dead-code spot checks,
model changes, identity changes, and remaining limits.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

trevon referenced this pull request

2026-09-22 21:13:12 +00:00

feat: resolve C/C++ and Rust calls across files and typed receivers #3

trevon merged commit 583c26588d into feat/cpp-rust-resolution

2026-09-22 21:15:02 +00:00

trevon referenced this pull request from a commit

2026-09-22 21:15:04 +00:00

Merge pull request 'feat: add brick_callers and brick_uncalled caller-status queries' (#4) from feat/caller-status into feat/cpp-rust-resolution