feat: add llm curation step

README.md

@@ -16,7 +16,7 @@ One tool, multiple projects, multiple LLM providers (Anthropic / OpenAI / Bedroc
 
 ## Install
 
-cliff-notes is published as a raw source repo (no compiled binary yet). Two ways to use it:
+cliff-notes is published as a raw Bun source repo (no compiled binary yet). Two ways to use it:
 
 ```sh
 # Ad-hoc, no install
@@ -37,14 +37,18 @@ model = "claude-sonnet-4-6"
 
 [project]
 name = "my-project"
-audience = "internal-devs"
-voice = "concise, technical, no marketing fluff"
+audience = "end-users of the application"
+voice = "clear, user-focused, concise, no marketing fluff"
 ```
 
 Then export the relevant API key (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`, or AWS credentials for Bedrock).
 
 If the project doesn't already have a `cliff.toml`, cliff-notes uses a bundled default. Override via `git_cliff.config = "path/to/cliff.toml"`.
 
+By default, cliff-notes runs a curation pass before rewriting entries. Commits sharing a PR number are grouped deterministically; remaining commits can be grouped or omitted by the model when the schema validates the full partition. Omission decisions use `project.audience` as free-form guidance, so external-user or operator notes can skip test-only and internal-only churn while maintainer notes can keep those changes when useful. Set `[curation] strategy = "by-pr-only"` for deterministic PR grouping only, or `"off"` for one bullet per commit. Use `--show-curation` to print the proposed groups and omissions before rewrite.
+
+GitHub PR enrichment reuses existing credentials: `GITHUB_TOKEN`, `GH_TOKEN`, or `gh auth token`. In GitHub Actions, `GITHUB_TOKEN` and `GITHUB_REPOSITORY` are enough for git-cliff PR enrichment; locally, an authenticated `gh` CLI is sufficient. Set `[github] enabled = false` to skip token/repo resolution.
+
 ## Use
 
 ```sh
@@ -62,6 +66,9 @@ cliff-notes --tag v1.2.3 --out release-notes.md
 
 # Skip the confirmation prompt
 cliff-notes --tag v1.2.3 --yes
+
+# Print grouping and omission decisions before rewrite
+cliff-notes --unreleased --dry-run --show-curation
 ```
 
 Either `--tag <version>` or `--unreleased` is required — cliff-notes does not infer version numbers.
@@ -84,7 +91,7 @@ Each generated section ends with an HTML comment containing the raw git-cliff en
 -->
 ```
 
-The block makes drift between raw commits and LLM rewrites diffable in code review. The marker version (`v1`) lets future cliff-notes re-render from the raw input without re-querying git.
+Grouped entries list every member commit in the audit block. Omitted commits do not appear as rendered bullets, but they are still listed in the audit block with the model's reason. The block makes drift between raw commits and LLM rewrites diffable in code review. The marker version (`v1`) lets future cliff-notes re-render from the raw input without re-querying git.
 
 ## Goreleaser integration (recipes)
 

cliff-notes.example.toml

@@ -17,8 +17,8 @@ model = "claude-sonnet-4-6"
 
 [project]
 name = "my-project"
-audience = "internal-devs"
-voice = "concise, technical, no marketing fluff"
+audience = "end-users of the application"
+voice = "clear, user-focused, concise, no marketing fluff"
 
 [prompt]
 # Optional. Appended to the baseline system prompt — use to teach the LLM
@@ -37,6 +37,25 @@ voice = "concise, technical, no marketing fluff"
 # own bundled default.
 # config = "cliff.toml"
 
+[curation]
+# strategy = "auto"        # default. PR prefilter + LLM curation on residual.
+# "by-pr-only" runs only deterministic PR grouping.
+# "off" keeps one changelog bullet per commit.
+# omit_plumbing = true     # LLM may suppress changes that are not meaningful
+# to the configured project.audience. Audience examples are guidance, not an
+# exhaustive list: external-user/operator notes can omit test-only changes and
+# developer-only churn, while maintainer notes can keep those when useful.
+# Omissions remain in the audit block.
+# min_group_size = 2
+# max_per_group = 5
+# max_index_gap = 15       # LLM cannot group commits more than 15 positions apart.
+# require_same_type = true # reject LLM groups that mix conventional commit groups.
+# cache = true             # content-hash cache for deterministic curation reruns.
+
+[github]
+# enabled = true           # reuse gh/GITHUB_TOKEN credentials for git-cliff PR enrichment.
+# repo = "owner/name"      # optional; default autodetected from env, gh, or origin.
+
 [output]
 changelog_file = "CHANGELOG.md"
 date_format = "%Y-%m-%d"

package.json

@@ -3,10 +3,10 @@
   "version": "0.1.0",
   "description": "Controlled LLM changelog generator (git-cliff + structured LLM)",
   "bin": {
-    "cliff-notes": "./dist/cli.js"
+    "cliff-notes": "./src/cli.ts"
   },
   "files": [
-    "dist",
+    "src",
     "cliff.toml",
     "cliff-notes.example.toml",
     "README.md"