search_graph already reads search_disable_dep_ranking, and the MCP schema names that config key. Add it to CBM_CONFIG_REGISTRY so config list/get guidance exposes the same setting users see in the API description.
Add a CLI registry regression test for the key, default, range, and dependency-ranking guidance. This does not change search behavior, MCP schemas, DB schema, or tool visibility.
Validated with make -f Makefile.cbm build/c/test-runner, CBM_ONLY_SUITE=cli ./build/c/test-runner, CBM_ONLY_SUITE=tool_consolidation ./build/c/test-runner, git diff --check, and bash scripts/check-source-safety.sh.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Use the Claude SessionStart matcher array length instead of the unrelated NUM_DIRS constant, and expose the helper with a Claude-specific name that matches the Gemini session hook API.
Add a focused CLI regression test that verifies startup, resume, clear, and compact SessionStart hooks are all installed and removed.
Validation: make -f Makefile.cbm build/c/test-runner; CBM_ONLY_SUITE=cli ./build/c/test-runner; bash scripts/check-source-safety.sh; git diff --check.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Add a skill_dirs_planned array to the install --plan receipt and route planned Claude Code skill directory writes there instead of instruction_files_planned.
This keeps the no-mutation install receipt accurate for agents reviewing planned changes before install. Real install behavior and hook upsert paths are unchanged.
Validation: make -f Makefile.cbm build/c/test-runner; CBM_ONLY_SUITE=cli ./build/c/test-runner; bash scripts/check-source-safety.sh; git diff --check; make -f Makefile.cbm cbm; HOME=/private/tmp/cbm-install-plan-smoke-20260630 ./build/c/codebase-memory-mcp install --plan.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Update generated agent instructions and SessionStart reminder text to use the same explicit auto-index conditions as the MCP schema: auto_index=true and under auto_index_limit.
Also describe search_code as text/regex source search instead of graph-augmented grep, avoiding an overclaim about behavior.
Validation: make -f Makefile.cbm build/c/test-runner; CBM_ONLY_SUITE=mcp ./build/c/test-runner; CBM_ONLY_SUITE=tool_consolidation ./build/c/test-runner; CBM_ONLY_SUITE=cli ./build/c/test-runner; bash scripts/check-source-safety.sh; git diff --check.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Tighten MCP and installed hook wording so auto-index behavior is described by its actual gates: auto_index=true and the repository remaining under auto_index_limit.
This keeps the streamlined/classic tool surfaces, hook matchers, and install behavior unchanged while making the descriptions more actionable for MCP clients and CLI hook users.
Validation: make -f Makefile.cbm build/c/test-runner; CBM_ONLY_SUITE=mcp ./build/c/test-runner; CBM_ONLY_SUITE=tool_consolidation ./build/c/test-runner; CBM_ONLY_SUITE=cli ./build/c/test-runner; bash scripts/check-source-safety.sh; git diff --check.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Port upstream OpenClaw MCP config behavior to the fork branch: OpenClaw now writes nested mcp.servers entries with enabled, command, and args instead of the generic mcpServers shape, while preserving existing config.
Also skip .claude-worktrees during discovery to avoid indexing stale duplicate worktree content alongside the existing .claude and .worktrees exclusions.
Verified: build/c/test-runner; build/c/codebase-memory-mcp; CBM_ONLY_SUITE=cli (107); CBM_ONLY_SUITE=discover (84).
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Use the bounded discover count for startup, first-use, and explicit path autoindexing so CBM_AUTO_INDEX_LIMIT cannot be bypassed by a later MCP handler path.
Move sync first-use indexing into resolve_project_store via a shared MCP-local helper, leaving REQUIRE_STORE responsible for joining an in-flight startup index and reporting missing stores. This avoids duplicate indexing attempts and duplicate over-limit warnings while preserving structured stderr logging.
Add an input-validation regression for over-limit explicit path projects and clarify the config help text to say the limit counts indexable files.
Verified: build/c/test-runner; build/c/codebase-memory-mcp; CBM_ONLY_SUITE=input_validation (44), mcp (113), discover (83), cli (104).
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Use the discover walker for MCP autoindex file-limit checks instead of shelling out to git ls-files, so non-git roots and ignored files follow the same filters as indexing.
Add typed effective config helpers so env-backed registry settings such as CBM_AUTO_INDEX_LIMIT and CBM_REINDEX_ON_STARTUP are honored consistently by MCP startup logic.
Repair the CBM_ONLY_SUITE discover allowlist and add regressions for bounded non-git discovery counts and env-over-db config priority.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Install, update, and uninstall cleanup paths are scoped to project indexes, but the old .db predicate also selected _config.db. Add a shared project-index predicate, exclude the persistent config store, and keep list/count/remove behavior consistent.
Verification: make -f Makefile.cbm build/c/test-runner; CBM_ONLY_SUITE=cli build/c/test-runner; make -B -f Makefile.cbm build/c/codebase-memory-mcp.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Remove brittle hardcoded tool-count wording from MCP hidden-tool and CLI config descriptions. Keep the concrete tool names and discovery instructions intact so maintainers and MCP clients see the same capabilities without stale numeric claims.
Verification: make -f Makefile.cbm build/c/test-runner build/c/codebase-memory-mcp; CBM_ONLY_SUITE=tool_consolidation build/c/test-runner; CBM_ONLY_SUITE=mcp build/c/test-runner; CBM_ONLY_SUITE=cli build/c/test-runner.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Default incremental_reindex to off so normal indexing uses the existing full atomic rebuild path until the disk incremental implementation avoids full-graph load, quality-pass, FTS rebuild, and dump work. This matches the measured behavior where small FastAPI edits can be slower through the current incremental path than a full rebuild.
Keep fast and always as explicit opt-ins for benchmarking and canary tests. Update pipeline tests that intentionally validate legacy incremental behavior to set incremental_reindex=always locally instead of relying on production defaults.
Validation: make -f Makefile.cbm build/c/test-runner cbm; CBM_ONLY_SUITE=pipeline build/c/test-runner: 212 passed; CBM_ONLY_SUITE=cli build/c/test-runner: 102 passed; git diff --check.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Add a config-backed incremental_reindex policy with default fast so full and moderate indexes rebuild atomically instead of entering the current disk incremental path, which still performs full-graph quality work and can be slower than a clean rebuild. Keep always for benchmarking and off for deterministic full rebuilds.
Harden the existing incremental path around allocation failures before mutation: classify, changed-file list, graph buffer creation, changed-path set, and deleted-path copies now fail closed or preserve existing nodes rather than dereferencing NULL or recording corrupt path arrays.
Reuse the existing pipeline nanosecond constant for file-hash mtimes and keep the dedicated incremental suite opted into incremental_reindex=always so its canaries continue exercising the legacy disk incremental implementation.
Validation: make -f Makefile.cbm build/c/test-runner cbm; CBM_ONLY_SUITE=pipeline build/c/test-runner: 212 passed; CBM_ONLY_SUITE=cli build/c/test-runner: 102 passed; CBM_ONLY_SUITE=mcp build/c/test-runner: 113 passed; CBM_ONLY_SUITE=tool_consolidation build/c/test-runner: 94 passed; git diff --check.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Attach the normal config store before one-shot CLI tool dispatch so CLI calls honor the same registry defaults and env/config overrides as the stdio MCP server. This keeps explicit path search auto-indexing enabled by default for user-facing CLI calls while embedded/no-config helpers remain manual unless CBM_AUTO_INDEX explicitly opts in.
Clarify tool/config/help strings to describe the real behavior: streamlined lists 5 core tools plus _hidden_tools discovery, get_architecture is available after reveal or classic mode, and default tools can auto-index the server CWD or explicit repository paths when enabled.
Validation:
- make -f Makefile.cbm cbm build/c/test-runner
- CBM_ONLY_SUITE=mcp build/c/test-runner: 113 passed
- CBM_ONLY_SUITE=tool_consolidation build/c/test-runner: 94 passed
- CBM_ONLY_SUITE=input_validation build/c/test-runner: 43 passed
- CBM_ONLY_SUITE=cli build/c/test-runner: 102 passed
- CLI smoke: isolated explicit repo path auto-indexed without CBM_AUTO_INDEX=true and returned cli_autoindex_sentinel
- CLI stdout/stderr split: JSON on stdout, logs on stderr
- git diff --check
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Document search_graph query as the BM25 path and make clear that graph filters and sort_by do not apply there.
Describe semantic_query as an array that appends separate semantic_results instead of implying it changes the normal result ranking.
Update installed agent guidance and hooks to prefer the streamlined get_code surface, avoid hyperbole, and describe auto-indexing as best-effort.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Limit install/update process termination to MCP servers running the exact target binary path instead of every process named codebase-memory-mcp. This prevents fake HOME or alternate-prefix installs from killing unrelated agent sessions.
Make Codex config upserts idempotent when the MCP table and SessionStart hook share config.toml. Preserve hook sentinels during MCP section replacement and repair stale duplicate/orphan CMM hook blocks on the next upsert.
Verification: make -f Makefile.cbm cbm; make -f Makefile.cbm build/c/test-runner; CBM_ONLY_SUITE=cli build/c/test-runner; fake-HOME install --force collapsed a corrupted Codex config to one MCP section and one SessionStart hook.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Update generated CLI skill text, packaged skill docs, and npm README examples for the current MCP surface: trace_path is the canonical trace tool, index_dependencies is present, and the streamlined/default surface should not be described as exactly 14 tools.
Replace unsupported search_graph direction examples with supported query_graph degree queries and fix stale query/result-limit wording in generated help. Keep measured performance/preprint claims unchanged.
Verified with: make -f Makefile.cbm cbm; make -f Makefile.cbm build/c/test-runner; CBM_ONLY_SUITE=cli build/c/test-runner; git diff --check.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Use trace_path as the single public call-tracing tool across classic and streamlined MCP surfaces, while exposing existing semantic_query and search_code mode parameters in the advertised schemas. Keep the streamlined surface compact, but ensure the hidden-tool reveal covers the full classic capability set without trace aliases.
Persist PageRank, LinkRank, and node_degree rows with direct VALUES inserts instead of per-row SELECT probes. The node/edge project is loaded once with the graph rows so full-scope dependency attribution stays identical to the previous behavior.
Add focused coverage for dependency project attribution, streamlined parameter contracts, and targeted suite allowlist entries so these checks cannot silently run zero tests.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Make index writes and artifact refreshes use verified temp files plus atomic replacement, preserving Windows replace semantics and avoiding fixed temp-path collisions.
Thread the remaining pipeline confidence knobs through config while keeping zero-valued defaults backward compatible with existing upstream behavior.
Tighten route-node gates and architecture route output so infra URL nodes stay queryable in the graph but do not pollute get_architecture route summaries; preserve code-created absolute URL routes for all language extractors.
Harden MCP and CLI path handling around cache paths, grep scratch files, update downloads, checksum validation, stale checks, and project DB discovery without writing protocol noise to stdout.
Add regression coverage for corrupt-DB recovery, concurrent atomic writes, tmp-prefixed project DBs, duplicate search result fields, route output filtering, config registry entries, and focused suite selectors.
Validated with: git diff --check; make -f Makefile.cbm build/c/test-runner; focused suites security, artifact, graph_buffer, sqlite_writer, store_search, store_bulk, store_pragmas, store_checkpoint, watcher, infrascan, pipeline, store_nodes, mcp, cli, parallel, tool_consolidation, incremental, and store_arch; make -f Makefile.cbm cbm; isolated dogfood index/query run.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Replace the consolidated search_code_graph mega-tool — which merged graph
search + source grep + cypher behind two nearly disjoint parameter
families — with the three existing focused tools as the DEFAULT surface:
search_graph (graph/structural), search_code (source grep), query_graph
(cypher). These are drawn from TOOLS[] (single schema source, DRY);
trace_call_path + get_code remain on the default surface. The
search_code_graph schema and its thin dispatch router are deleted; the
underlying handlers (handle_search_graph/search_code/query_graph) stay.
The _hidden_tools progressive-disclosure hint now lists 11 hidden tools
(was 14). Repointed three user-facing hint strings (handle_trace_call_path
missing-name and function-not-found; two unknown-tool path hints) from the
deleted tool to search_graph.
Tests updated to the new 5-tool surface:
- tests/test_mcp.c: mcp_tools_list + server_handle_tools_list assert the
5 default tools present and search_code_graph absent.
- tests/test_tool_consolidation.c: renamed streamlined_mode_shows_3_tools
-> ..._5_default_tools; repurposed the two search_code_graph dispatch
tests into search_graph_dispatch / query_graph_dispatch (query_graph
handler accepts both cypher= and query=); revised
schema_compact_documented_as_graph_only (the graph-vs-source compact
collision it warned about no longer exists post-split).
- tests/test_input_validation.c: remapped 16 dispatch calls by mode --
cypher -> query_graph, source-grep -> search_code, graph/structural ->
search_graph.
- tests/test_depindex.c: tool_index_dependencies_listed asserts
search_graph (was search_code_graph).
- src/cli/cli.c: six config-description strings (search_limit, tool_mode,
context_injection, compact, default_sort_by, default_include_dependencies).
Back-compat: blast radius is in-repo only (0 references in Go internal/cmd,
docs, or the CGO wrapper). External callers of search_code_graph must
switch to search_graph / search_code / query_graph, which are strictly
more capable. query_graph keeps its query= backward-compat alias.
Context -- of the three section-4 fixes in the merge plan, only section 4b
was a real change: section 4c (function<->class edges) was already
implemented (commit b950bf2 added the missing MEMBER_OF reverse-edge
coverage), and section 4a (passive resources) was already mitigated by
inject_context_once plus the get_graph_schema/get_architecture/index_status
tools -- both verified against code, not assumed.
Gate: 5981 passed / 2 failed (the 2 are the deferred B1 store-layer
incremental failures, unrelated to section 4b).
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Deep --no-ff merge of upstream DeusData/main (597 commits, merge-base 988d975)
into fork api-consolidation (HEAD 3018b22), producing a superset that combines
upstream breadth with the fork's surviving depth.
Upstream additions integrated: semantic/embedding search (vendored/nomic),
minhash/SimHash, cross-repo + cross-file LSP + route + pkgmap passes, artifact
import/export, path_alias, git_context, hook_augment, Windows support
(compat_fs/compat_thread), the UI layer, and ~140 expanded tree-sitter grammars.
Fork depth retained: PageRank/LinkRank ranking (src/pagerank/; sort_by=
relevance/linkrank/degree/calls), dependency indexing (src/depindex/;
CBM_INDEX_MODE=DEP), token reduction, and HTTP-endpoint discovery
(src/pipeline/httplink.c, pass_httplinks.c).
Conflict resolution (27 regions across the 39-file intersection):
- src/mcp/mcp.c: kept the fork superset handler pattern (resolve_project_store +
expand_project_param + inject_context_once) as the single path; UNIONed the
tool dispatch table; emit status:"indexed", excluded[], cross_service trace.
- src/store/store.c: upstream base + fork pagerank/linkrank/node_degree tables
and sort_by plumbing; cbm_store_node_degree now excludes INHERITS
(WHERE type != 'INHERITS'); removed the duplicate cbm_store_get_db.
- src/pipeline/pipeline.c: took upstream run_extraction_phase; grafted the fork
passes (httplinks, normalize, flush_store dump branch); call
cbm_pipeline_set_committed_counts before cbm_gbuf_dump_to_sqlite.
- internal/cbm/lang_specs.c: 158 entries combining the fork section_node_types
with upstream ts_factory/embedded_imports; wired 13 orphaned type-arrays.
- src/depindex/depindex.h: renamed CBM_PATH_MAX -> CBM_DEP_PATH_MAX (it collided
with constants.h enum CBM_PATH_MAX=1024) + 11 call sites in depindex.c.
- Makefile.cbm: upstream grammar/nomic units + fork CGO targets; added
httplink.c + pass_httplinks.c to PIPELINE_SRCS; removed mongoose NOSAN
machinery (mongoose was deleted upstream).
- httplinks restored (the merge silently deleted them because the fork had left
them unchanged since the base): httplink.c/.h, pass_httplinks.c,
tests/test_httplink.c; registered RUN_SUITE(httplink) in tests/test_main.c.
- CBM_INDEX_MODE enum unified to FULL=0, MODERATE=1, FAST=2, DEP=3.
Completeness (dual-parent check): 0 upstream files dropped from the merge tree;
6 fork files dropped, all intentional (upstream-superseded): CHANGELOG.md,
vendored/mimalloc/src/{arena-abandon,segment-map,segment}.c,
vendored/mongoose/{mongoose.c,mongoose.h}. Fork features (pagerank, depindex,
token_reduction, httplink) verified present and registered in tests/test_main.c.
src/pipeline/pass_parallel.c has an identical function set to upstream (49=49,
none dropped); the large line-deltas vs upstream are within-function divergence
of the fork extractor, not dropped functions.
Gate (run manually; hooks skip on merge commits): build green
(make -f Makefile.cbm cbm, exit 0); full C test suite 5980 passed / 2 failed /
0 crashes. The fork's pre-existing ASan crash in build_project_list_error_srv is
fixed by upstream's mcp.c rewrite, so the merge is strictly better than the fork
baseline on the crash axis.
Known issue, tracked separately (NOT a merge defect): 2 nondeterministic
failures in tests/test_incremental.c — incr_db_deleted_recovery:772 (diff_pct)
and incr_accuracy_vs_full:803 (full_nodes vs incr_nodes). Cause: the fork's
extractor (internal/cbm/extract_*.c) and parallel-resolve Route-node creation
are not thread-safe under the concurrent use of upstream's parallel pipeline
(run_parallel_pipeline), producing nondeterministic route/infra HANDLES edges
(~234k edge swing between runs). Sequential extraction (CBM_WORKERS=1) is
deterministic and correct. Root cause, bisect, and repro are documented in
notes/2026-06-24-fork-vs-upstream-divergence-assessment.md (section 8f); the fix
is deferred to a dedicated thread-safety audit (the make test-tsan target is
currently a stub and needs wiring).
Under -D_GNU_SOURCE with optimization, glibc's string.h macros make
strstr() on a `const char *` (with a constant needle) yield a
`const char *`. gcc 15 then rejects assigning that to `char *` with
-Werror=discarded-qualifiers, breaking the build:
src/cli/cli.c:1503: error: initialization discards 'const'
qualifier from pointer target type [-Werror=discarded-qualifiers]
Minimal repro (gcc 15.2):
const char *c = "hello";
char *p = strstr(c, "ell"); // gcc -D_GNU_SOURCE -O2 -Werror=discarded-qualifiers
codex_hook_strip only reads through begin/end/cut (offset math and
byte comparisons, never writes), so making them `const char *` is the
correct fix and preserves the function's existing const contract
(`content` is already const). The malloc'd output buffer stays char *.
No behavior change; build succeeds on gcc 15.2.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: Jeremy Maserang <gothuncc@gothaunty.com>
* fix(cli): copy and sign the built binary on install instead of re-signing the stale target
`install` (especially with --force) only ad-hoc-signed whatever already sat at
~/.local/bin/codebase-memory-mcp and never copied the binary the operator ran
install from. An operator who built from source and ran `install -y --force`
was silently left with the OLD binary at the target — running stale code while
believing they had upgraded (#472).
Install now mirrors `update`: detect the running binary, copy it to the
canonical target (skipping the copy when we already ARE the target, which would
otherwise truncate the running file), make it executable, ad-hoc sign the
TARGET, and point all agent configs at the target. A different pre-existing
binary is only replaced with --force or interactive confirmation; --dry-run
reports the intended swap without touching anything.
The copy+sign also resolves the secondary report: a hand-copied clang build is
linker-signed (flags=0x20002) and gets Killed:9 when spawned by an MCP host —
re-signing ad-hoc (flags=0x2) makes it launchable, which install now does for
the placed binary.
Adds cbm_copy_binary_to_target() (with a same-file guard) plus regression tests
covering the binary swap and the self-copy guard.
Closes#472
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: Shane McCarron <shane.mccarron@corvexconnect.com>
* fix(cli): address QA round 1
Extend cli_install_same_file_guard_issue472 with a symlink case: two distinct
path strings resolving to the same inode (which a non-canonical
cbm_detect_self_path vs the hardcoded target can produce) must also be detected
as same-file and skipped, not truncated. POSIX-only (#ifndef _WIN32).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: Shane McCarron <shane.mccarron@corvexconnect.com>
* test(cli): guard install copy exec-bit assertion to POSIX
cli_install_copies_binary_to_target_issue472 asserted the executable bit on the
copied target, but cbm_copy_binary_to_target only chmods on POSIX (the exec bit
is not meaningful on Windows, and MinGW stat() derives it from the file
extension). Wrap the stat/S_IXUSR assertion in #ifndef _WIN32 so test-windows
passes; the content and creation assertions still run on all platforms.
Refs #472
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: Shane McCarron <shane.mccarron@corvexconnect.com>
---------
Signed-off-by: Shane McCarron <shane.mccarron@corvexconnect.com>
Co-authored-by: Shane McCarron <shane.mccarron@corvexconnect.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
The standard linux release binary dynamically links glibc 2.38+ and
GLIBCXX_3.4.32, so it fails to start on Debian 11, Ubuntu 20.04/22.04,
RHEL/Rocky 8/9, Amazon Linux 2, etc. — yet install.sh, the npm and PyPI
wrappers, and the binary's own self-update all fetched it by default,
contradicting the "single static binary" promise.
Point every linux install + self-update path at the fully-static
"-portable" asset (gcc -static), which has no glibc floor. macOS/Windows
are unaffected and unchanged.
- install.sh, pkg/npm/install.js, pkg/pypi _cli.py: select -portable on linux
- src/cli/cli.c: self-update download URL AND checksum archive name both use
-portable on linux (they must match or the update fails checksum verify)
- scripts/smoke-test.sh: assert linux self-update targets the -portable asset
- scripts/ci/check-glibc-compat.sh: new guard — runs the binary inside
debian:bullseye (glibc 2.31) and asserts it starts
- _smoke.yml: run the guard on the portable binary in smoke-linux-portable
Reproduced: standard binary -> "GLIBC_2.38 not found" on glibc 2.31;
portable binary runs cleanly.
Antigravity support targeted the pre-unification layout, so it never actually
configured the CLI:
- detection probed ~/.gemini/antigravity/ — the CLI now installs under
~/.gemini/antigravity-cli/ (brain/, mcp/, settings.json)
- MCP config was written to ~/.gemini/antigravity/mcp_config.json — Antigravity
reads the SHARED ~/.gemini/config/mcp_config.json (mcpServers, command/args)
- the SessionStart hook + AGENTS.md targeted the stale dir
Point detection at ~/.gemini/antigravity-cli/, write the MCP server to the
shared ~/.gemini/config/mcp_config.json (creating ~/.gemini/config if needed),
and place AGENTS.md + the SessionStart reminder under ~/.gemini/antigravity-cli/.
Install, uninstall, install --plan, the detection test, and the README table
updated.
Codex, Gemini CLI, and Antigravity all gained lifecycle-hook support, so give
them the same non-blocking SessionStart reminder Claude Code has — stdout is
injected as session context, nudging the agent to use codebase-memory-mcp
graph tools before grep.
- Codex: append a sentinel-delimited [[hooks.SessionStart]] block to
~/.codex/config.toml (idempotent upsert, preserves other content, removable).
- Gemini CLI: SessionStart hook in ~/.gemini/settings.json, reusing the shared
JSON hook upsert (alongside the existing BeforeTool reminder).
- Antigravity: same JSON hook in ~/.gemini/antigravity/settings.json (it shares
Gemini hook semantics).
The reminder command is written to be valid both as a TOML single-quoted
literal and a JSON string (no single quotes, no newlines). Wired into install,
uninstall, and the install --plan receipt. remove_hooks_json now prunes an
emptied event key so removal leaves no stale "<Event>": [] cruft. README
Multi-Agent table updated. Tests cover Codex upsert/idempotency/removal and
Gemini parity.
An installer that configures MCP entries, instruction files, Skills, and hooks
across many agents should let callers see what it will mutate before it does.
`--dry-run` already lists planned writes in human-readable form; this adds the
machine-readable `agent.install.plan.v1` JSON receipt the issue asked for.
`install --plan` runs the real install dispatch in record-only mode — a global
g_install_plan recorder that each write-site function appends to (at the same
point it would write) while all mutations are disabled — then emits JSON with
agents_detected, config_files_planned, instruction_files_planned,
hooks_planned, writes_started:false, network_after_install:false, and
next_safe_command. Because the plan is recorded on the actual install code
path, it cannot drift from real behavior. No config is written, no index
deleted, no network used.
cbm_build_install_plan_json is exposed for testing; the test asserts the
receipt content AND that building it creates no config files.
cbm_ensure_path appended `export PATH="<dir>:$PATH"` to every shell rc,
including config.fish — which is a syntax error in fish and broke users
fish config. Detect a .fish target and emit `fish_add_path <dir>` instead
(idempotent, prepends only if absent). POSIX shells are unchanged.
The symlink-attack vector (predictable /tmp/cbm-code-discovery-gate-$PPID)
was removed in c29e6d5 when the blocking bash gate became a stateless
compiled augmenter writing to stdout only. Add a regression guard so a
predictable temp/state file can never be reintroduced: assert the emitted
gate shim contains no /tmp path and no $PPID, and delegates to hook-augment.
Export cbm_install_hook_gate_script (was static) so the security property
is directly testable.
Cursor was never in the editor-detection table, so install/update only
ever wrote the VS Code config and silently skipped ~/.cursor/mcp.json,
even when Cursor was present. Uninstall removed the Cursor entry but a
subsequent install never re-created it.
Detect ~/.cursor/ as the Cursor agent, register the MCP server in
~/.cursor/mcp.json (mcpServers format, same as Windsurf/Gemini) on
install, remove it on uninstall, and list Cursor in detected agents.
Agent CLIs on Windows are installed as .cmd/.ps1/.exe shims (e.g. opencode
via mise/npm), but find_in_path only probed the bare name, so they weren't
detected. Try .exe/.cmd/.bat/.ps1 per PATH entry on Windows.
Relates to #221 (opencode not detected). The install-time taskkill 'eq'
error from the same report is addressed separately — cbm_kill_other_instances
now uses _spawnvp with an argv array rather than a shell string.
Route Claude Code config paths (skills, .mcp.json, .claude.json,
settings.json, hook scripts) and agent detection through
CLAUDE_CONFIG_DIR-aware helpers, falling back to ~/.claude. Hook command
strings written to settings.json keep the legacy tilde form when the env
var is unset, so existing configs stay portable across HOME values. Prints
a one-line migration nudge when CLAUDE_CONFIG_DIR is set and a legacy
~/.claude tree still exists.
Adds cli_detect_agents_finds_claude_via_env and isolates CLAUDE_CONFIG_DIR
in the existing detection tests so the runner env can't leak in.
Distilled from #321 onto current main (adapts to the v0.7.0 non-blocking
augmenter hook signature, which the original branch predated). Closes#320.
Format drift had accumulated across recent commits (the dry-run lint enforces clang-format-20, which had not run on these commits yet). No functional changes — whitespace/layout only, verified by a clean cppcheck + clang-format + NOLINT lint-ci pass.
The previous PreToolUse hook gated Grep/Glob/Read/Search with 'exit 2'
on the first call per session, which broke Claude Code's
read-before-edit invariant (issue #362) and could deny tool calls under
upgrade/missing-binary failure modes.
Replace it with a structurally non-blocking augmenter:
- New 'codebase-memory-mcp hook-augment' subcommand reads the hook JSON
from stdin and, for Grep/Glob, queries search_graph (in-process, no
shell) and emits hookSpecificOutput.additionalContext. Every failure
path (no project, short token, missing binary, slow query, timeout)
exits 0 with no stdout — the hook physically cannot block a tool call.
- 300 ms SIGALRM/_exit(0) in-process deadline; 5 s settings.json timeout
backstop. Output is written exactly once at the very end, so a
mid-work timeout yields a clean no-op (never partial JSON).
- Matcher narrowed to 'Grep|Glob' (Read explicitly excluded) for Claude;
Gemini matcher narrowed to 'google_search|grep_search' (excludes
read_file) for the same reason.
- The installed shim is a thin wrapper that delegates to the binary;
legacy filename 'cbm-code-discovery-gate' is kept so existing
settings.json entries upgrade with zero migration. Installer refuses
to embed binary paths containing a double quote (shim injection
defense).
- Per-agent 'old matchers' lists let upsert/remove clean up historical
matcher strings during upgrade.
- Smoke tests (8d/8e/8l) updated to assert the new behavior and
regress-test against re-introducing Read in the matcher or 'exit 2'
in the shim.
- Session reminder text updated: 'always Read a file before editing it'
replaces the prior 'fall back to Read only for text content'.
(cherry picked from commit f72c8e68c4d91e52911a569a967ad782ce5472b2)
Detect ~/.kiro/ directory, auto-configure MCP at ~/.kiro/settings/mcp.json
using standard mcpServers format. Adds install, uninstall, detection, and test.
Based on #174 by dLo999, rebased onto refactored install infrastructure.
Co-Authored-By: Dustin Obrecht <dustin@kurtnoble.com>
check_already_latest() exits early before printing dry-run info when
the running version matches the latest release. Skip this check in
dry-run mode so the user sees what would happen.
Lead with what it does, then why to enable/disable with multiple examples.
Use colons not em-dashes. Cover non-code sessions, token-metered environments,
CI pipelines, and scripted use as disable reasons. Include exact commands.
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
Adds boolean config key context_injection (default true, env CBM_CONTEXT_INJECTION)
that controls whether inject_context_once embeds the _context header in the first
tool response.
WHAT _context contains: node/edge counts, all node labels with counts, edge types
with counts, PageRank stats, and detected ecosystem (Go/Python/Rust/etc.).
WHEN TO DISABLE (context_injection=false):
- Scripted/programmatic use: extra JSON adds noise to parsed output
- CI pipelines: token cost is metered, model needs no schema context
- Model has explicit codebase instructions via system prompt
- Benchmarking: removes schema-query overhead from latency measurements
WHEN TO KEEP ENABLED (default):
- Interactive AI sessions benefit from automatic situational awareness
- Saves 2-3 round-trips otherwise needed for get_architecture/get_graph_schema
Usage:
codebase-memory-mcp config set context_injection false # per-install
CBM_CONTEXT_INJECTION=false codebase-memory-mcp # per-session
Implementation:
- Config check placed BEFORE setting context_injected flag so toggling
mid-session works (re-enable without server restart)
- session_project is still always included regardless of this setting
- cbm_config_get_bool reads CBM_CONTEXT_INJECTION env var automatically
via cbm_config_get_effective (the standard env-override path)
TDD: config_context_injection_disabled, config_context_injection_enabled_by_default
Tests: 2274 passed, 0 failed
Signed-off-by: Andrew Hundt <ATHundt@gmail.com>
All hardcoded ~/.cache/codebase-memory-mcp paths now route through
cbm_resolve_cache_dir() in platform.c. Priority: CBM_CACHE_DIR env
var > default ~/.cache/codebase-memory-mcp.
Added db_path_for_project() helper in http_server.c to eliminate 5
copies of the same path construction pattern.
Fixes#154
Co-Authored-By: dLo999 <dLo999@users.noreply.github.com>
Check GitHub releases/latest redirect header before downloading.
Saves bandwidth and avoids unnecessary index rebuilds.
- "Already up to date" when version matches or is ahead
- --force flag to bypass the check
- Graceful degradation when network is unavailable
- Uses same curl dependency as the download itself
Fixes#142
Co-Authored-By: dLo999 <dLo999@users.noreply.github.com>
PATH uses ; on Windows, not : — strtok with : splits C:\Users at
the drive letter colon, producing garbage paths. Also skip S_IXUSR
check on Windows since stat() doesn't set Unix permission bits.
Fixes#159
Co-Authored-By: slvnlrt <slvnlrt@users.noreply.github.com>
Replaces hardcoded ~/.local/bin/ with OS-native runtime detection:
- Windows: GetModuleFileNameA()
- macOS: _NSGetExecutablePath()
- Linux: readlink("/proc/self/exe")
Falls back to ~/.local/bin/ if detection fails.
Fixes install.ps1 placing binary in $LOCALAPPDATA/Programs/ while
configs point to non-existent ~/.local/bin/ path.
Fixes#158
Co-Authored-By: slvnlrt <slvnlrt@users.noreply.github.com>
KiloCode is a VS Code extension — uses the same base config dir.
Apply same platform-aware paths as Zed/VS Code: macOS uses
$HOME/Library/Application Support/Code/, Windows uses %APPDATA%/Code/.
Merge codebase-memory-exploring, codebase-memory-tracing,
codebase-memory-quality, codebase-memory-reference into single
codebase-memory skill with decision matrix, workflows, and gotchas.
Old skill directories are cleaned up automatically during install.
Factual corrections vs original PR: trace_call_path -> trace_path,
removed non-existent tools (read_file, list_directory) and edge types
(OVERRIDE, USAGE, FILE_CHANGES_WITH), removed risk_labels parameter.
Co-Authored-By: gdilla <gdilla@users.noreply.github.com>
Andrew Hundt