subscription-research-agent

name	subscription-research-agent
description	Use when an agent needs deep research from subscription sources, local research workspaces, archived evidence, evidence briefs, source-backed research memos, hypothesis tracking, or multi-step synthesis.

Subscription Research Agent

Overview

Use this skill to orchestrate local-first research workflows around subscription information channels. Deterministic tools prepare evidence packages; the active Agent writes final analysis from that evidence when the user asks.

Workflow Selection

• For RSS article discovery, daily news, key information, or quick digest generation, use rss-ai-digest.
• For source quality and registry maintenance, use rss-source-curator.
• For local research workspace setup, run subscription-research init.
• For archiving RSS results into the research workspace, run subscription-research ingest rss.
• For optional full-text evidence enrichment, run subscription-research content fetch or subscription-research content summarize after RSS ingest and before generating the evidence brief.
• For source-backed research context, run subscription-research brief evidence.
• For historical source health review, failed-feed analysis, or registry maintenance, use rss-source-curator.
• For a research daily, deep-dive memo, or source-backed synthesis, generate or read an evidence brief first, then follow references/daily-report.md.
• For workspace structure, read references/research-workspace.md.
• For evidence brief fields, read references/evidence-brief.md.
• For daily report structure, read references/daily-report.md.
• For the daily report quality gate, read references/daily-report-quality.md.

Runtime Command

Examples use subscription-research as a compatibility command. Before running commands, resolve the runtime command:

• When DIGESTPILOT_BACKEND_URL is available, product actions such as evidence preparation, daily registration, or WebUI-backed research tasks may use the Backend task API first, then read the resulting workspace output.
• If the Backend API is unavailable, keep using the local runtime command fallback below. Do not require WebUI, cloud, or a hosted service in generic Agent environments.
• Use DIGESTPILOT_RUNTIME_CMD when the environment provides a custom runtime command.
• Use digestpilot-runtime when it is available on PATH.
• Use subscription-research when only the legacy compatibility command is available on PATH.
• From a repository checkout without a linked command, replace subscription-research with node packages/research-cli/dist/src/cli.js.
• Run node scripts/doctor.mjs from the repository root to diagnose local runtime setup.
• If no runtime command is available but this repository or plugin package includes runtime source, run node scripts/bootstrap-runtime.mjs, then retry the command.

Core Commands

Initialize a workspace:

subscription-research init --workspace research-workspace

Archive RSS evidence with the default Node RSS runtime:

subscription-research ingest rss \
  --workspace research-workspace \
  --registry feeds.json \
  --since 24h \
  --keywords "llm,agent,rag,evals,inference" \
  --should-keywords "benchmark,reliability,architecture" \
  --exclude-keywords "webinar,coupon,sponsor,hiring,job,press release" \
  --min-score 7

Optionally enrich archived articles with readable full text:

subscription-research content fetch \
  --workspace research-workspace \
  --since 7d \
  --min-score 7 \
  --limit 20 \
  --timeout 20

Generate an evidence brief:

subscription-research brief evidence \
  --workspace research-workspace \
  --question "What changed in LLM evals this week?" \
  --since 7d \
  --must-keywords "llm,evals" \
  --must-keyword-mode all \
  --should-keywords "benchmark,reliability,agent"

For broad research dailies, prefer topic words as --should-keywords or use --must-keyword-mode any so the brief does not require every AI keyword to appear in one article.

Source-health history can be used as research context only when it materially changes evidence confidence. For source maintenance reports or failed-feed triage, route to rss-source-curator.

Boundaries

This skill does not promise a final research conclusion from deterministic tooling alone. Treat generated briefs as evidence packages. The Agent should cite evidence items when writing a memo or report.

Research dailies and memos are Agent-written synthesis artifacts. The CLI prepares evidence; the Agent owns judgment, recommended reading order, and follow-up questions.

When a research daily will be consumed by the DigestPilot WebUI, preserve the compact daily handoff fields described in references/daily-report.md: daily metadata, ordered sections, and item title/link/source/author/summary.

Full-text enrichment is optional and local-first. Use it when RSS summaries are too thin for research synthesis. Do not require it for ordinary RSS daily news or quick key-information lookups.

When article content has already been enriched for a DigestPilot detail page, reuse the same extracted content fields as research evidence instead of fetching the page again. Preserve extraction status and summary_source so the Agent can distinguish cleaned body excerpts from RSS summaries.

Do not use this skill for ordinary RSS daily news or quick key-information lookups. Those should stay in rss-ai-digest. Use rss-source-curator for feed maintenance output.