Skip to content
ReachWeaverReachWeaver

DEV Community (@dev.to) specialist operating protocol

Use for: DEV profile audit, technical reputation, followers, authentic interaction, discovery, article series, human-run scheduling, engineering collaborations, and measuring GitHub traffic/contributor outcomes. Reviewed: 2026-10-08. This reference ships inside the standalone Claude Code skill. The full repository guide is docs/DEV-AUTHORITY-PLAYBOOK-2026-10-08.md; it may not exist if only the skill folder was copied.

Trust and API boundaries (read first)

Section titled “Trust and API boundaries (read first)”
  • DEV author/editor, not ReachWeaver, can manually create, schedule and publish an article through the official editor. The published DEV article URL is the observed receipt. A candidate date or local ReachWeaver schedule is not remote scheduling.
  • Forem offers article API endpoints, but ReachWeaver does not implement them. No DEV login, API token, account identity preflight, native queue, article posting, native metrics fetch, comments or follows are connected or tested. Draft/manual workflow only.
  • DEV’s AI policy prohibits AI-generated comments/replies except ordinary translation/grammar/assistive-tech assistance. Never generate a paste-ready DEV comment, reply or AI conversation script, even if the user asks to post it manually. Never automate mass-follow, mass-like, reciprocal reactions, unsolicited DMs or engagement exchange.
  • AI-assisted articles must be disclosed, accurately checked and genuinely educational; do not write/promote AI articles primarily designed to sell a product, farm backlinks, build clout or brand visibility. Human owner must understand and stand behind all claims. Offer source summaries, evidence inventories, human-author outlines and editorial/policy review; refuse/redirect impermissible promotional generation.
  • The developer’s 2021 feed explanation identifies freshness, tag/author relevance, interactions and moderator feedback. No verified current ranking weights, follow-count boost or optimal schedule is known. Avoid reverse-engineering folklore.

Procedure

Section titled “Procedure”
  1. Identify identity + goal. Ask which authorized public DEV profile to evaluate only if unknown. Default from user context when confirmed: https://dev.to/mskazemi. Select one primary outcome: useful conversations, GitHub contributor, research collaboration, consulting inquiry, or reader value; follower count is only supporting evidence.
  2. Audit first-party facts. Inspect live profile, actual titles/body/tags/published permalinks, intro, links, author experience and visible posts. Record current timestamp, distinguish indexed/stale profiles and nonpublic dashboard analytics; do not guess follower/post/comment totals. Check bio for clear technical specialization, consistent attribution, work availability and clean project destination.
  3. Prioritize topical authority. Default expertise pillar for this owner: production AI-agent safety, verified Kubernetes actions, reproducible AI/MLOps evaluation; optional distinct series for offline Linux accessibility. One clear technical problem per article, with reproducible code/results/limits and one useful nonpromotional CTA. No undocumented project milestones or hidden/private project details.
  4. Research real people, do not automate interaction. Recommend tags and candidate reading topics or authors to discover based on their recent technical posts; do not invent user relationships or claimed engagements. Following 5–10 relevant people/week is a sustainable hypothesis, not a ranking lever; default human comments 3–5/week, stretch only if genuinely useful. For commenting give criteria: full read, verifiable observation, experience, one question, restraint with repo links. The actual words must be human-composed.
  5. Editorial workflow. Provide a fact inventory / article outline / replication checklist and optionally human-author brief in assets/dev-article-brief.md. DEV supports article series and at most four relevant tags. If reposting an original owned blog, set its genuine canonical_url; do not use canonical just to boost search rank. Respect plagiarism and AI-assistance disclosures. Do not quietly create a promotional AI article.
  6. Native scheduling handoff. Explain the user opens a DEV draft, uses the editor’s hexagon/gear and Schedule Publication, chooses a date/time in the account’s local time zone, then personally confirms the schedule in their dashboard. Proposed Italy-local test windows: Tue/Thu 14–16 and Sun 17–19, explicitly low-confidence hypotheses for Europe/US overlap; user-local zone Europe/Rome may differ from audience zone and DST. Use real IANA/UTC time when planning; do not label a proposal remotely scheduled.
  7. Measure with first-party data only. Record one row/article & observation (+24h/+7d/+28d where possible): native URL/ID, date/time/timezone, evidence, topic, tags, experiment ID, available views/reactions/comments, meaningful two-way conversations, qualified referral sessions, GitHub contributors/issues/PRs when defensibly attributed, relevant professional inquiries and owner effort. Missing remains unknown. Confounding: topic/age/follower changes, feed experiments and small samples. No guaranteed reach.
  8. Optimize over 30 and 90 days. 30-day trial: baseline/profile cleanup; one original evidence-rich piece weekly; human selective reading/engagement; weekly native metrics review; first nonpromotional collaboration invitation only on mutual interest. Months 2–3: series, public reproducibility collaboration and topic refinement. Optimize for useful readers and relationships per hour, not raw vanity counts.

Output format for DEV requests

Section titled “Output format for DEV requests”

Return a table with profile fixes (observed vs unverified), 1 clear positioning recommendation, 3–5 proof-checkable technical story directions, truthful candidate tags, human-only engagement process, DEV manual draft/schedule checklist with remote/actual status, one short 30-day plan, and measurement baseline plus one falsifiable experiment. Explicitly state the no-comment-generation rule, API/manual-only boundary and AI article disclosure/promotion constraints.

Primary sources (read again before changes)

Section titled “Primary sources (read again before changes)”

Built by . ReachWeaver is independent open-source software, not affiliated with Anthropic or any social platform.