maint: port verified community fixes onto current main

[tirth8205]

Summary

Ports four verified community fixes onto current main without restoring their stale/conflicting branch history:

• fix: extend default ignore patterns and fix nested path matching #92: distinguish safe-anywhere dependency directories from root-only output names, while preserving fix: expand trailing slash ignore patterns #583 trailing-slash ignore support.
• fix(install): write opencode MCP config in official schema #530: use OpenCode's official opencode.jsonc / opencode.json, top-level mcp, and local command-array schema.
• fix: use lowercase-hyphen skill names #562: use lowercase-hyphen skill names in all generated and bundled skills, including the static review-changes file missing from the stale diff.
• docs: quote pip extras install examples #625: quote pip extras with double quotes across all current user docs, fixing zsh without breaking Windows cmd.exe.

Each source contributor is retained through a Co-Authored-By trailer on the corresponding commit.

Verification

• Full suite: 1,476 passed, 2 xpassed
• Incremental/ignore suite: 62 passed
• Skills/install suite: 169 passed
• Documentation regressions: 2 passed
• Ruff on every changed Python file: clean
• git diff --check: clean

The locally installed repo-wide Ruff version also reports 44 pre-existing findings in untouched files; this branch introduces none and cleans the two touched test modules.

Fixes #91.
Fixes #561.

Source contributions: #92, #530, #562, #625.

[github-actions]

code-review-graph review

Overall risk: 0.55 (MEDIUM) — 27 changed function(s)/class(es), 0 affected flow(s), 11 test gap(s)

Risk-scored changes

Risk	Level	Symbol	Location	Tested
0.55	medium	code_review_graph/skills.py::install_platform_configs	code_review_graph/skills.py:404	no
0.40	medium	code_review_graph/incremental.py::_load_ignore_patterns	code_review_graph/incremental.py:361	no
0.40	medium	code_review_graph/incremental.py::_should_ignore	code_review_graph/incremental.py:388	no
0.35	low	code_review_graph/skills.py::_build_server_entry	code_review_graph/skills.py:241	no
0.35	low	code_review_graph/skills.py::_warn_legacy_opencode_config	code_review_graph/skills.py:261	no
0.30	low	code_review_graph/skills.py::_opencode_config_path	code_review_graph/skills.py:35	no
0.30	low	tests/test_documentation.py::test_pip_extra_examples_use_cross_shell_double_quotes	tests/test_documentation.py:31	(test)
0.30	low	tests/test_documentation.py::test_current_user_docs_have_no_unquoted_pip_extras	tests/test_documentation.py:40	(test)
0.30	low	tests/test_incremental.py::TestDataDirRegistry	tests/test_incremental.py:376	no
0.30	low	tests/test_incremental.py::TestDataDirRegistry.test_registry_fallback_to_env_var	tests/test_incremental.py:420	(test)

Test gaps

• code_review_graph/incremental.py::_load_ignore_patterns (code_review_graph/incremental.py:361)
• code_review_graph/incremental.py::_should_ignore (code_review_graph/incremental.py:388)
• code_review_graph/skills.py::_opencode_config_path (code_review_graph/skills.py:35)
• code_review_graph/skills.py::_build_server_entry (code_review_graph/skills.py:241)
• code_review_graph/skills.py::_warn_legacy_opencode_config (code_review_graph/skills.py:261)
• ...and 6 more without direct tests

Token savings: this graph-backed report used ~99,823 fewer tokens (~95%) than reading every changed file in full (estimated, chars/4 approximation).

---

Powered by code-review-graph — local-first analysis; no code leaves the CI runner.