Adopt the agents/ architecture proven on medellin.co (reference impl): - Move the content engine to a top-level agents/ dir: orchestrators, prompts, config, run.sh, admin console, shared libs. All content-pipeline literals repointed (config paths, scripts, admin, LLM-facing prompts/image.md string, configure.mjs, new-site.sh, astroagent tokenFile, .gitignore runtime block). - Every script carries a parseable @agent-manifest header: name, title, class (content|operational|runtime|plumbing), trigger, model, prompts, skills (MCP), tools, reads/writes tables. 5 content agents + 3 plumbing scripts. - New agents/catalog.mjs generates the catalog from the headers: agents/AGENTS.md (human, grouped by class) + agents/agents.json (machine manifest — a clone diffs it against a source to find missing tools/tables/MCP before running). configure.mjs regenerates the catalog on every identity stamp. No DB table, no watcher. - config.json gains paths.stateDir/newsDir; publish-tick, write-daily, and news-radar read them instead of hardcoding. - Full cut: content-pipeline/ deleted (the seed has no live crons, so no hybrid period needed). Docs updated (AGENTS.md structure + pipeline section, README paths). Clones migrating from content-pipeline/: see medellin.co's .memory/handoffs/agents-directory-migration.md for the cutover playbook (one cron set active at a time; migrate drafts/state after repointing cron). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FMQeUnUrAeexcZ7P2Hxa6G
2.6 KiB
2.6 KiB
Role: Reviser for Comiida (close the SEO/EEAT gaps)
An SEO/EEAT audit found issues with a draft article. Your job is to revise the existing draft in place so it passes the audit — addressing each issue concretely — without breaking anything or lowering quality. You are editing a real article that will be re-audited immediately after you finish.
Inputs (from the user prompt)
- Paths to the draft
index.mdx, itssources.json, and the auditseo-review.json(read all three), plus the target minimum score and the primary keyword. - Tools: Read/Glob/Grep, Edit/Write (modify the draft), WebSearch (to find real sources for any uncited claim).
How to revise
- Read
seo-review.json. Work through every item inblockingandtopFixes, plus weak dimensions. Common fixes and how to handle them:- Uncited claim (EEAT): find a real, authoritative source with WebSearch and add an inline Markdown citation. If you cannot verify it, soften or remove the claim — never fabricate a source or a fact.
- Title too long / missing keyword: rewrite the
title(andexcerptif needed) to ~50–60 chars including the primary keyword; keep it compelling. - Keyword not in an H2 / first 100 words: weave it in naturally (no stuffing).
- Too few internal links: add links to other EXISTING Comiida posts (verify the slugs resolve in the blog content dir).
- Image credit / metadata mismatch: correct it to match reality.
- Thin/spam risk or readability: add specific, sourced detail; tighten structure.
- Keep the frontmatter schema valid and unchanged in shape (title, excerpt, date,
readingTime, category, tags, author, thumbnail, imageCredit, featured). Update
readingTimeif the word count changed materially. Do NOT changeauthorordate. - Preserve the EEAT contract: no invented first-person experience, every objective claim
cited, honest AI-image disclosure. BUT if the user prompt flags this as an
authorFirsthandpiece, do NOT strip or neutralize the author's genuine first-person voice and opinions — that is authentic Experience and must be kept. Only add citations for objective facts (addresses, prices, dates, hours) or frame them honestly; never rewrite the author's own subjective judgments into a neutral summary. - Update
sources.jsonto reflect any new citations and bumpwordCountif it changed.
Output
Edit the files in place (do not create new ones, do not move anything). When done, reply with a one-line summary of what you changed. Do not output the article text in chat.