--- name: scout.hugen.tokyo description: scout.hugen.tokyo provides keyword search across six developer-focused content platforms: Hacker News, Lobste.rs, Stack Overflow/StackExchange, Dev.to, PyPI, and GitLab. Each skill returns structured metadata (titles, URLs, scores, authors, dates) for matching results. All skills are single-shot query endpoints — no streaming, no full-content retrieval, no write operations. host: scout.hugen.tokyo --- # scout.hugen.tokyo scout.hugen.tokyo is a read-only search aggregator targeting developer research workflows. It serves agents that need to discover, compare, or monitor content across the major developer community platforms from a single host. Its distinguishing characteristic is breadth across complementary surfaces: link aggregators (HN, Lobste.rs), Q&A (Stack Overflow), blog content (Dev.to), package registries (PyPI), and code hosting (GitLab). It does not cover GitHub, npm, crates.io, or any non-developer platforms. ## When to use this host Use scout.hugen.tokyo when an agent needs keyword-based discovery or monitoring across HN, Lobste.rs, Stack Overflow, Dev.to, PyPI, or GitLab in a single host. It is well-suited for research tasks: topic landscape scans, dependency vetting, and community sentiment gathering. Do not use it for: fetching a specific item by ID (use the native platform APIs directly); retrieving full answer or article body text (only metadata is returned); GitHub repository search (use the GitHub Search API); npm, crates.io, or other non-PyPI package registries; real-time or streaming front-page feeds; security audits, CVE lookups, or dependency graph traversal; or any write, comment, or vote operations on any platform. ## Capabilities ### Developer Link Aggregators Searches curated tech link aggregators to surface community-ranked stories, discussions, and articles on a topic. Useful for gauging community interest and finding high-signal links. - **`search-hacker-news`** — Searches Hacker News stories, comments, and polls via Algolia and returns titles, URLs, points, comment counts, authors, and dates for matching results. - **`search-lobsters-posts`** — Searches Lobste.rs, a curated tech link aggregator, and returns matching posts with scores, comment counts, authors, tags, and timestamps. ### Q&A and Technical Articles Retrieves question metadata from Stack Overflow/StackExchange and article metadata from Dev.to, covering both problem-solution Q&A and longer-form technical blog content. - **`search-stackoverflow-questions`** — Searches Stack Overflow (or any StackExchange site) by keyword and returns questions with scores, answer counts, view counts, tags, author info, and direct URLs. - **`search-devto-articles`** — Searches Dev.to (Forem) for technical articles and returns title, description, URL, author, reactions, comments, reading time, tags, and publication date for each result. ### Package and Project Discovery Looks up Python packages on PyPI and public projects on GitLab, returning version, license, author, and repository metadata needed to evaluate or audit a dependency or tool. - **`search-pypi-packages`** — Looks up Python packages on PyPI by name (with hyphenated, underscored, and py- prefix variants) and returns version, summary, license, author, Python requirements, release count, and links. - **`search-gitlab-projects`** — Searches GitLab.com for public projects and returns name, description, stars, forks, topics, default branch, and last activity date for each result. ## Workflows ### Topic Landscape Scan *Use when an agent needs to build a broad picture of community discussion, articles, and Q&A around a specific technology or topic across multiple developer platforms.* 1. **`search-hacker-news`** — Retrieve top HN stories and discussions for the topic to gauge community interest and find high-signal links. 2. **`search-lobsters-posts`** — Retrieve Lobste.rs posts on the same topic to capture curated tech community perspective, which often differs from HN. 3. **`search-stackoverflow-questions`** — Retrieve Stack Overflow questions to identify common problems and pain points practitioners encounter with the topic. 4. **`search-devto-articles`** — Retrieve Dev.to articles to surface tutorial and explainer content written by practitioners. ### Python Dependency Research *Use when an agent needs to evaluate a Python library — resolving its canonical package name, checking metadata, and then finding community discussion and known issues.* 1. **`search-pypi-packages`** — Resolve the package name to its PyPI entry, capturing latest version, license, author, and homepage URL. 2. **`search-stackoverflow-questions`** — Search Stack Overflow for questions about the package to surface common usage problems and gotchas. 3. **`search-hacker-news`** — Search HN for discussions about the package to find community sentiment, launch threads, or notable criticism. ### GitLab Project Evaluation *Use when an agent needs to discover a GitLab-hosted project and then assess community awareness or find usage guidance beyond the repository metadata.* 1. **`search-gitlab-projects`** — Find the project on GitLab and capture stars, forks, topics, last activity, and description. 2. **`search-stackoverflow-questions`** — Search Stack Overflow for questions referencing the project to gauge adoption and surface integration issues. 3. **`search-devto-articles`** — Search Dev.to for articles about the project to find tutorials or community write-ups. ## Skill reference ### `search-hacker-news` **Scout HN – Hacker News Search** — Searches Hacker News stories, comments, and polls via Algolia and returns titles, URLs, points, comment counts, authors, and dates for matching results. *Use when:* Use when an agent needs to discover or analyze Hacker News posts on a topic, retrieve top-scoring stories, filter by content type (story, comment, poll, show_hn, ask_hn), or sort results by relevance or date. *Not for:* Do not use for real-time live HN front-page feeds or streaming updates; this is a single-shot search query. Not suitable for fetching a specific HN item by ID — use the HN API directly for that. **Inputs:** - `q` (string, required) — Search query string to match against HN content. - `sort` (string) — Sort order for results: 'relevance' (default) or 'date' for chronological ordering. - `per_page` (integer) — Number of results to return, between 1 and 50. Defaults to 10. - `tag` (string) — Filter results by content type. One of: story, comment, poll, show_hn, ask_hn. **Returns:** Returns success=true with a data array of up to 50 HN items (id, title, url, hn_url, author, points, comments, date, type), plus source, query, count, and elapsed_ms fields. **Example:** `GET https://scout.hugen.tokyo/scout/hn?q=AI+agents&sort=relevance&per_page=10&tag=story` --- ### `search-gitlab-projects` **GitLab Project Search** — Searches GitLab.com for public projects and returns name, description, stars, forks, topics, default branch, and last activity date for each result. *Use when:* Use when an agent needs to discover or evaluate public GitLab-hosted projects by keyword, especially for projects not mirrored on GitHub such as enterprise OSS, Linux tooling, or GitLab-native repositories. *Not for:* Do not use for security audits, CI/CD pipeline analysis, or code quality reports — this endpoint returns only project metadata. Do not use to search GitHub-hosted repositories; use a GitHub search API instead. **Inputs:** - `q` (string, required) — Search query string for matching GitLab project names and descriptions. - `per_page` (integer) — Number of results to return per page. Minimum 1, maximum 20, default 10. - `sort` (string) — Sort order for results. Allowed values: stars, updated, name. Default is stars. **Returns:** Returns success=true and a data array of up to 20 GitLab project objects each containing id, name, full_name, description, url, stars, forks, topics, default_branch, last_activity, and created timestamps. **Example:** `GET https://scout.hugen.tokyo/scout/gitlab?q=kubernetes&per_page=10&sort=stars` --- ### `search-pypi-packages` **PyPI Scout** — Looks up Python packages on PyPI by name (with hyphenated, underscored, and py- prefix variants) and returns version, summary, license, author, Python requirements, release count, and links. *Use when:* Use when an agent needs to resolve a Python package name to its PyPI metadata, including latest version, license, author, homepage URL, and upload date, before installing or auditing a dependency. *Not for:* Do not use for dependency graph traversal or vulnerability/CVE scanning; those capabilities are not returned by this endpoint. Not suitable for streaming package index updates. **Inputs:** - `q` (string, required) — Package name or approximate name to look up on PyPI. The endpoint tries hyphenated, underscored, and py- prefix variants automatically. - `per_page` (integer) — Number of results to return, between 1 and 20. Defaults to 5. **Returns:** Returns success=true with a data array of PyPI package objects each containing name, version, summary, license, author, pypi_url, homepage, requires_python, release_count, and latest_upload timestamp. **Example:** `GET https://scout.hugen.tokyo/scout/pypi?q=fastapi&per_page=5` --- ### `search-stackoverflow-questions` **StackScout Stack Overflow Search** — Searches Stack Overflow (or any StackExchange site) by keyword and returns questions with scores, answer counts, view counts, tags, author info, and direct URLs. *Use when:* Use when an agent needs to find Stack Overflow or StackExchange Q&A content matching a keyword or topic, such as looking up programming solutions, retrieving top-voted answers, or gathering metadata about community questions. *Not for:* Do not use for fetching a specific question by ID or retrieving full answer body text; this endpoint returns question-level metadata only, not full answer content. **Inputs:** - `q` (string, required) — Search query string to match against Stack Overflow questions. - `per_page` (integer) — Number of results to return per page. Minimum 1, maximum 50. Defaults to 10. - `sort` (string) — Sort order for results. One of: relevance, votes, activity, creation. Defaults to relevance. - `site` (string) — StackExchange site to search. Defaults to stackoverflow. **Returns:** Returns success=true with a data array of up to 10 question objects each containing id, title, url, score, answers, views, is_answered, tags, author, author_reputation, and date, plus source, query, count, and elapsed_ms. **Example:** `GET https://scout.hugen.tokyo/scout/stackoverflow?q=async+await&per_page=10&sort=votes&site=stackoverflow` --- ### `search-lobsters-posts` **Lobsters Scout Feed** — Searches Lobste.rs, a curated tech link aggregator, and returns matching posts with scores, comment counts, authors, tags, and timestamps. *Use when:* Use when an agent needs to find recent or trending tech articles from Lobste.rs by keyword, such as discovering community-curated posts on a specific language, tool, or topic. *Not for:* Do not use for Hacker News results; use a separate HN search endpoint instead. Not suitable for fetching full article content or comment threads — only post metadata is returned. **Inputs:** - `q` (string, required) — Search query string to match against Lobste.rs posts. - `per_page` (integer) — Number of results to return per page. Minimum 1, maximum 25, default 10. - `sort` (string) — Sort order for results. Accepted values: 'hot' (default) or 'newest'. **Returns:** Returns success=true with an array of up to 25 Lobste.rs posts, each containing id, title, url, lobsters_url, score, comments, author, tags, and date, plus source, query, count, and elapsed_ms. **Example:** `GET https://scout.hugen.tokyo/scout/lobsters?q=rust&per_page=10&sort=hot` --- ### `search-devto-articles` **Dev.to Article Search** — Searches Dev.to (Forem) for technical articles and returns title, description, URL, author, reactions, comments, reading time, tags, and publication date for each result. *Use when:* Use when an agent needs to discover or monitor Dev.to technical blog posts on a given topic, retrieve article metadata for analysis, or surface trending or recent content from the Dev.to community. *Not for:* Do not use for fetching a single known article by ID or URL; this is a keyword search endpoint only. Not suitable for real-time streaming feeds or non-Dev.to platforms. **Inputs:** - `q` (string, required) — Search query string to find matching Dev.to articles. - `per_page` (integer) — Number of results to return per page. Minimum 1, maximum 30, default 10. - `sort` (string) — Sort order for results. One of: relevance, latest, top. Default is relevance. **Returns:** Returns success=true and a data array of up to 10 article objects, each with id, title, description, URL, author name/username, reactions, comments, reading_time, tags, and date. **Example:** `GET https://scout.hugen.tokyo/scout/devto?q=typescript&per_page=10&sort=relevance` ---