Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add SkillMedev/skills --skill engagement-reply-draftergit clone --depth 1 https://github.com/SkillMedev/skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/skillmedev/skills/engagement-reply-drafter)<a href="https://agentmods.dev/skills/skillmedev/skills/engagement-reply-drafter"><img src="https://agentmods.dev/badge/skills/skillmedev/skills/engagement-reply-drafter/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/skillmedev/skills/engagement-reply-drafter"><img src="https://agentmods.dev/badge/skills/skillmedev/skills/engagement-reply-drafter.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00097 | $0.00871 |
| Opus 5 | $0.00048 | $0.00436 |
| Sonnet 5 | $0.00019 | $0.00174 |
| Haiku 4.5 | $0.00010 | $0.00087 |
Grade A, and why
Engagement Reply Drafter scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
Copies of this mod
1 near-identical copy found in the catalogue:
- Engagement Reply Drafter — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Engagement Reply Drafter
Turn a comment section or DM inbox into specific, human, on-brand replies that grow the thread without sounding like a bot.
Workflow
- Triage each item into one lane: praise, question, complaint/bug, criticism, troll/bad-faith, or spam/abuse. The lane sets the tone; a mismatched lane produces the wrong reply. Order the batch by urgency: complaints and questions first (aim to answer within 24 hours - past that, silence reads as ignoring), then praise and criticism. Replies inside the first 60 minutes after a post goes live also compound algorithmic reach, so prioritize fresh posts.
- Read what the person actually said and name a concrete detail from it. Never open with a generic platitude ("Thanks so much!").
- Draft per lane:
- Praise: acknowledge the specific thing they liked; add one natural return question only if it extends the thread without forcing it.
- Question: answer it directly first, then optionally invite more.
- Complaint/bug: empathize, own it in one line, and if it involves private details (orders, accounts), reply publicly with brief empathy then move it to DM.
- Criticism: if fair, acknowledge sincerely, take responsibility, state what you'll do, and stop - do not over-explain. If bad-faith, write a brief unbothered reply or recommend no reply.
- Spam/abuse: do not engage; recommend hide or block.
- Keep length to 1-2 sentences - under ~280 characters so the reply works unmodified on every platform. Social comments are not email.
- Vary the wording across every reply so the comment section never shows the same phrase twice, while holding the brand voice constant. Mirror given voice traits; default to warm, concise, human.
- Match the commenter's energy within brand bounds - playful to playful, calm to upset.
- Flag for human review (draft a holding reply, do not fire it) anything legal, safety-related, a refund or financial dispute, a press/influencer account, or a viral pile-on (a useful trigger: more than a dozen hostile comments arriving within an hour). Mark it as needing review.
- Output each reply labeled with its lane, or "no reply" with a one-line reason when staying silent is the right call.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 46 lines · 97 tokens per session scan A c45a646c4701
Engagement Reply Drafter is a skill published in the GitHub repository SkillMedev/skills (12 stars, last pushed 2mo ago), licensed MIT. It adds 97 tokens to every session and 871 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
webinar-funnel-builder
Designs a complete webinar funnel - registration page promise, reminder email sequence timed against typical 35-45 percent show-up rates, webinar structure with a 70 percent teach / 30 percent offer split and the pivot script, replay window and urgency rules, and the post-webinar email sequence - with a funnel-metrics…
YouTube Script Writer
Writes retention-engineered long-form YouTube scripts on a hook/body/payoff architecture - a first-30-seconds hook that survives the retention cliff, a beat-sheet body with mini-hooks and bridges, timed pattern interrupts, and a payoff that over-delivers - with B-roll and editing cues marked inline. Use when someone…
Show Notes Writer
Produces complete podcast show notes from a transcript or episode outline - a standalone summary block, timestamped chapters that name what happens, screenshot-worthy key takeaways, guest bio, curated links, and keyword placement for search. Use when someone says "write show notes for this episode", "turn this…
Policy Brief
Writes a concise, evidence-based policy brief - problem, options compared on consistent criteria, and a recommendation - that a decision-maker can act on in minutes. Use when someone asks "write a policy brief", "summarize the evidence and options for this decision", "brief the minister/board/council on X", or needs a…
Systematic Review
Runs a PRISMA-compliant systematic review from PICO question and pre-registered protocol through search strings, two-stage screening, risk-of-bias appraisal, and GRADE-rated synthesis. Use when someone asks "run a systematic review", "build my search strategy and inclusion criteria", "screen these studies against a…
Thumbnail Concept
Produces three distinct high-CTR YouTube thumbnail concepts per video - each a full spec with focal point, precise facial-expression direction, text overlay copy, color and background direction, and two paired title variants. Use when someone says "I need a thumbnail idea", "why is my CTR low", "help me pick a…