perform-forge-review: Add AI-driven forge review skill

Add a skill for AI-assisted code review workflow on GitHub, GitLab, or
Forgejo. The key design principle is human oversight: the AI builds
review comments in a local JSONL file which the human can inspect and
edit before submission. Reviews are submitted as pending/draft, allowing
final edits in the forge UI.

Includes scripts/reviewtool, a Python 3 CLI with no external dependencies
that handles checkout, comment building, and submission to various forges.

Migrated from bootc-dev/infra PR #64.

Assisted-by: OpenCode (Claude Sonnet 4)
Signed-off-by: Colin Walters <walters@verbum.org>

Author: cgwalters

perform-forge-review/SKILL.md

@@ -0,0 +1,71 @@
+---
+name: perform-forge-review
+description: Create AI-assisted code reviews on GitHub, GitLab, or Forgejo. Use when asked to review a PR/MR, analyze code changes, or provide review feedback.
+---
+
+# Perform Forge Review
+
+Create code reviews on GitHub, GitLab, or Forgejo with human oversight.
+
+## Workflow
+
+Use `scripts/reviewtool` for all operations. It requires Python 3 with no
+external dependencies.
+
+### 1. Check out the PR
+
+```bash
+scripts/reviewtool checkout 123
+```
+
+This checks out the PR branch and shows the diff. For GitLab/Forgejo, set
+the appropriate environment variables first.
+
+### 2. Review the code
+
+Read the files, understand the changes. Use `git diff` and `git log` as needed.
+
+### 3. Build comments
+
+Start a review, then add comments:
+
+```bash
+scripts/reviewtool start --pr 123 --body "Assisted-by: OpenCode (Claude Sonnet 4)
+
+AI-generated review. Comments prefixed with AI: are unedited."
+
+scripts/reviewtool add --pr 123 \
+  --file src/lib.rs --line 42 --match "fn process_data" \
+  --body "AI: *Important*: Missing error handling"
+```
+
+The `--match` flag validates the line content to prevent misplaced comments.
+
+### 4. Submit
+
+```bash
+scripts/reviewtool submit --pr 123
+```
+
+The review is created as pending/draft. The human reviews in the forge UI
+and submits when satisfied.
+
+## Comment conventions
+
+**Prefixes:**
+- `AI: ` — unedited AI output
+- `@ai: ` — human question for AI to process
+- No prefix — human reviewed/edited
+
+**Priority markers:**
+- `*Important*:` — must resolve before merge
+- (none) — normal suggestion
+- `(low)` — minor nit
+
+## Review body
+
+Do not summarize the PR changes. The body should contain:
+- Attribution (required)
+- Concerns not tied to specific lines (optional)
+
+Avoid positive-only inline comments — they create noise.