git:20260725.09f86fd to git:20260908.e0fa3af

21 added, 18 removed. Audit A to A.

AUTONOMY_LEVEL: aggressive

Solo OSS maintainer running with `bypassPermissions`. Atomic-ship convention applies (commit + push + tag + push tag + `gh release create` in one turn) — see `feedback_claudemd_ship_from_main_atomic.md`. §5.1 Never-downgrade set binds as written in core §5.1 (single source; the verbatim list was deduped here 2026-07-25 — it drifts).

+

+ ## claude-mem-lite — persistent memory

+ PreToolUse hooks already run `mem_recall` for past lessons before Read/Edit/Write. The calls worth making proactively:

+ | When | Call |

+ |------|------|

+ | Before Edit/Write | hook already recalled; if a `#NN` lesson was injected, cite `#NN` next time you produce user-visible text (citing = adopting the feedback; uncited lessons decay) |

+ | After fixing a non-trivial bug | `mem_save(type="bugfix", lesson_learned="<root cause + fix>", importance=2)` |

+ | After a non-obvious architecture decision | `mem_save(type="decision", lesson_learned="<constraint + tradeoff>")` |

+ | Deferring to a future session | `mem_defer({title, priority:1|2|3, detail})`; when fixed, add `closes_deferred=[N]` to `mem_save` |

+ | Looking up past work / history | `mem_search "keywords"` · `mem_recent` · `mem_timeline` |

+ Path cost is round-trips, not milliseconds: the PreToolUse hook above already recalls (0 calls) — prefer it. For an explicit query, if these `mem_*` tools are deferred behind ToolSearch this session, the Bash CLI `claude-mem-lite` is one call vs two (ToolSearch + call); the MCP server instructions carry the absolute path to use when it is not on PATH.

+ Full tool + CLI tables, citation/decay rules, and save discipline → `.claude/plugin_claude_mem_lite.md`

+

## Code Graph (repo-wide AST index)

AST + FTS + vector index of the whole repo — prefer over multi-round Grep/Read for

structural queries (LSP only sees open files; this sees everything). Fastest path = Bash CLI:

| Intent | Command |

|--------|---------|

| Who calls X / what X calls | `code-graph-mcp callgraph X` |

| Impact before editing a fn | `code-graph-mcp impact X` |

| Unfamiliar dir / module | `code-graph-mcp overview <dir>` |

| Symbol source / signature | `code-graph-mcp show X` |

| Concept search (no exact name) | `code-graph-mcp search "…"` (vector: MCP `semantic_code_search`) |

| grep + AST context | `code-graph-mcp grep "pat" [paths] [-t lang] [-g glob] [-c]` |

+ Not on PATH? A plugin-only install keeps its own copy — same commands, run

+ `~/.cache/code-graph/bin/code-graph-mcp` (or `npm i -g @sdsrs/code-graph` once).

Still use Grep for literal strings/regex in non-code files; still Read files you'll edit.

Full command + MCP-tool table: `.claude/plugin_code_graph_mcp.md`

-

- ## claude-mem-lite — persistent memory

- PreToolUse hooks already run `mem_recall` for past lessons before Read/Edit/Write. The calls worth making proactively:

- | When | Call |

- |------|------|

- | Before Edit/Write | hook already recalled; if a `#NN` lesson was injected, cite `#NN` next time you produce user-visible text (citing = adopting the feedback; uncited lessons decay) |

- | After fixing a non-trivial bug | `mem_save(type="bugfix", lesson_learned="<root cause + fix>", importance=2)` |

- | After a non-obvious architecture decision | `mem_save(type="decision", lesson_learned="<constraint + tradeoff>")` |

- | Deferring to a future session | `mem_defer({title, priority:1|2|3, detail})`; when fixed, add `closes_deferred=[N]` to `mem_save` |

- | Looking up past work / history | `mem_search "keywords"` · `mem_recent` · `mem_timeline` |

- Path cost is round-trips, not milliseconds: the PreToolUse hook above already recalls (0 calls) — prefer it. For an explicit query, if these `mem_*` tools are deferred behind ToolSearch this session, the Bash CLI (exact path in the detail doc) is one call vs two (ToolSearch + call).

- Full tool + CLI tables, citation/decay rules, and save discipline → `.claude/plugin_claude_mem_lite.md`

-