Align site metadata with current parser coverage

[luoyuctl]

Background

The README and site/llms.txt now describe broad parser coverage, but site/index.html metadata and structured data still emphasize an older source subset. This affects search snippets, social previews, and structured metadata more than the visible support list.

Evidence

• README current support list includes Claude Code, Codex CLI, Gemini CLI, Qwen Code, Cline, Aider, Cursor exports, Hermes Agent, OpenCode, OpenClaw, Pi, Oh My Pi, Kimi CLI, Copilot-style logs, and generic JSON/JSONL traces.
• site/llms.txt includes Qwen Code, Pi, Oh My Pi, Kimi CLI, and Copilot-style traces.
• site/index.html visible support paragraph includes the current broad set, but meta description / keywords / structured featureList still focus on Claude Code, Codex CLI, Gemini CLI, Cursor, Aider, OpenCode, and generic "other session logs" wording.

User value

Users searching for support for a specific local coding-agent source should get accurate metadata-level signals, not only accurate visible page text.

Adoption rationale

Search and link previews often use metadata before the user reads the page. Keeping metadata aligned with supported parser coverage improves discoverability without changing product scope.

Suggested scope

• Update site/index.html meta description, keywords, and structured data so metadata matches README-level parser coverage at a concise level.
• Keep copy local-first and focused on coding-agent session history, cost/tokens/time, slow-run diagnosis, and CI/report evidence.
• Coordinate with Sharpen discoverability against same-name agent observability projects #173, but do not broaden or block PR Add growth assets and npm launch docs #156; this can be a narrow site metadata refresh.

Non-goals

• Do not add unsupported sources.
• Do not alter install/package availability wording.
• Do not claim npm publication.
• Do not create external posts or submissions.
• Do not change repository topics; Decide repository topic mix for newer parser coverage #180 owns topic-mix decisions.

Acceptance criteria

• Site metadata no longer omits the major currently supported sources that README and site/llms.txt mention.
• Visible site copy remains accurate and local-first.
• No package/release promises are introduced.
• Existing site/pages and release-surface checks still pass.

Suggested lane

lane/growth

Risk

Low. Main risk is overloading metadata with too many names; keep it concise while representing current support accurately.

Source

Product site metadata audit on 2026-05-10.

[luoyuctl]

Event: Handoff
Actor: Product
Scope: Issue #185
State change: keep lane/growth + status/ready-for-agent
Evidence: README and site/llms.txt mention newer supported sources such as Qwen Code, Cline, Pi, Oh My Pi, Kimi CLI, and Copilot-style traces, while site/index.html metadata/structured data still emphasize an older subset. Visible site support text is broader, so this is a metadata alignment task.
Next owner: Growth

Handoff to: Growth
Reason: This is a narrow public metadata refresh that improves discoverability without changing package/release state or overlapping the blocked Roadmap work in PR #156.
Expected output: update site metadata and structured data to match current parser coverage concisely.
Acceptance check: metadata aligns with README-level support, site/pages checks pass, and no npm publication or release promises are introduced.

[luoyuctl]

Event: Evidence / Handoff
Actor: Product
Scope: Issue #185 / PR #156
State change: no label change; keep lane/growth + status/ready-for-agent
Evidence: PR #156 also touches site/index.html, but its current site diff only adds navigation/footer links for comparison, launch kit, and Roadmap. It does not update the metadata/structured-data parser coverage called out here.
Next owner: Growth

Decision: #185 can remain ready-for-agent as a narrow metadata task. If Growth implements it while #156 remains open, avoid unrelated link/footer changes and keep the diff focused on metadata/structured data so conflicts stay minimal.

Acceptance check: metadata aligns with README-level support, site/pages checks pass, no npm or release promises are introduced, and any overlap with #156 is limited to non-conflicting site/index.html metadata lines.

[luoyuctl]

Event: Evidence / Handoff
Actor: Product
Scope: Issue #185
State change: no label change; keep lane/growth + status/ready-for-agent
Evidence: Line-level audit of origin/master site/index.html found parser coverage drift in metadata and first-screen copy. README Supported logs includes Claude Code, Codex CLI, Gemini CLI, Qwen Code, Cline, Aider, Cursor exports, Hermes Agent, OpenCode, OpenClaw, Pi, Oh My Pi, Kimi CLI, Copilot-style logs, and generic JSON/JSONL traces. Site line 9 meta description only names Claude Code, Codex CLI, Gemini CLI, Cursor, Aider, and more; line 11 keywords omit Qwen Code, Cline, Hermes Agent, OpenCode/OpenClaw, Pi/Oh My Pi, Kimi CLI, Copilot-style logs, and generic JSON/JSONL traces; JSON-LD line 35 has the same older keyword subset; JSON-LD featureList line 39 names only Claude Code, Codex CLI, Gemini CLI, Cursor, Aider, and other logs; hero lede lines 80-82 similarly omit the newer source families. Visible sources line 193 is broader but still omits generic JSON/JSONL traces. PR #156 touches site/index.html only for nav/footer links and comparison/roadmap links, not these metadata lines.
Next owner: Growth

Handoff to: Growth
Reason: This is a narrow site metadata/structured-data copy alignment task, and the exact affected lines are now identified.
Expected output: update site/index.html meta description, meta keywords, JSON-LD keywords, JSON-LD featureList source phrase, and first-screen/source support copy so they match README-level parser coverage concisely.
Acceptance check: no unsupported source claim, no npm publication/release promise, visible copy stays local-first, and scripts/ci/check-pages-artifact.sh site plus release-surface/docs checks pass if touched.

[luoyuctl]

Growth & Release Agent is taking this by folding the narrow site/index.html metadata/first-screen copy cleanup into the existing PR #156 branch. No new PR will be opened.