index.json

{
  "$schema": "https://raw.githubusercontent.com/5dive-ai/loops/main/schema.json",
  "version": 1,
  "loops": [
    {
      "slug": "ci-analyst",
      "jobTitle": "Competitive Intelligence Analyst",
      "tagline": "Watches every competitor and the field, catches what changed, and writes a digest before it matters.",
      "agent": [
        "community"
      ],
      "character": "dude",
      "skills": [
        "deep-research",
        "compile-knowledge"
      ],
      "trigger": {
        "type": "schedule",
        "cron": "0 */4 * * *",
        "label": "Every 4h · daily digest"
      },
      "starterPrompt": "Scan our competitor set and the field for the last interval — launches, pricing, funding, notable chatter. Update the watchlist and, once a day, write a concise sourced briefing of what changed and what it means for us, then post it to the team.",
      "model": "claude-opus-4-8",
      "effort": "high",
      "tags": [
        "research",
        "market-intel",
        "competitive"
      ],
      "proof": {
        "receipts": "worth-knowing.md intel log; daily intel briefs; competitors_watchlist.md (updated multiple times/day)",
        "blogPost": "https://blog.5dive.com/blog/replace-ci-analyst-with-ai"
      },
      "feature": true,
      "rarity": "rare"
    },
    {
      "slug": "intel-brief",
      "jobTitle": "Intel Brief — Analyst to Writer relay",
      "tagline": "Two teammates in one loop: an analyst extracts what changed, hands off, and a writer turns it into a one-paragraph team brief.",
      "agent": [
        "community",
        "marketing"
      ],
      "character": "dude → theo",
      "agents": [
        {
          "role": "analyst",
          "persona": "dude",
          "summary": "From the last interval's activity, pick the 3 shifts most worth the team's attention. Return only 3 short bullets."
        },
        {
          "role": "writer",
          "persona": "theo",
          "handoff": true,
          "summary": "Turn the analyst's 3 bullets into a single tight paragraph (max 60 words), lead with what matters most."
        }
      ],
      "skills": [],
      "trigger": {
        "type": "schedule",
        "cron": "0 */4 * * *",
        "label": "Every 4h"
      },
      "starterPrompt": "An analyst extracts the few shifts that matter from the last interval and returns three bullets. A writer receives them via {{previous_output}} and turns them into a one-paragraph team brief. One scheduled loop, two roles, one structured handoff.",
      "model": "claude-opus-4-8",
      "effort": "medium",
      "tags": [
        "research",
        "market-intel",
        "multi-agent",
        "relay"
      ],
      "proof": {
        "receipts": "the reference multi-agent loop for SPEC §4; every run signs a chain receipt"
      },
      "feature": true,
      "rarity": "epic"
    },
    {
      "slug": "content-marketer",
      "jobTitle": "Content Marketer / Blog Writer",
      "tagline": "Picks a topic from fresh intel, researches it, drafts in brand voice, fact-checks, and publishes.",
      "agent": [
        "marketing"
      ],
      "character": "theo",
      "skills": [
        "deep-research",
        "humanizer",
        "copywriting"
      ],
      "trigger": {
        "type": "schedule",
        "cron": "0 9 * * *",
        "label": "Daily"
      },
      "starterPrompt": "Pick a topic from the latest intel, research it deeply, draft a post in our brand voice, fact-check and humanize it, then publish.",
      "model": "claude-opus-4-8",
      "effort": "high",
      "tags": [
        "marketing",
        "content",
        "blog"
      ],
      "proof": {
        "receipts": "50+ live posts at blog.5dive.com"
      },
      "feature": false,
      "rarity": "common"
    },
    {
      "slug": "social-media-manager",
      "jobTitle": "Social Media Manager",
      "tagline": "Turns each post into IG story / FB / Telegram-channel posts, formats the creative, and schedules them.",
      "agent": [
        "marketing",
        "creative"
      ],
      "character": "theo",
      "skills": [
        "social",
        "video-use",
        "frontend-design",
        "ad-creative"
      ],
      "trigger": {
        "type": "schedule",
        "cron": "0 10 * * *",
        "label": "On publish · daily"
      },
      "starterPrompt": "For each newly published post, produce IG story / FB / Telegram-channel variants, format the creative, and schedule them.",
      "model": "claude-sonnet-4-6",
      "effort": "medium",
      "tags": [
        "marketing",
        "social"
      ],
      "proof": {
        "receipts": "live FB page; @5dive.ai IG; @ai5dive Telegram channel"
      },
      "feature": false,
      "rarity": "common"
    },
    {
      "slug": "community-manager",
      "jobTitle": "Community Manager",
      "tagline": "Watches comments across YouTube/IG/FB, surfaces what needs a reply, and drafts responses.",
      "agent": [
        "community"
      ],
      "character": "dude",
      "skills": [
        "social",
        "notify-user"
      ],
      "trigger": {
        "type": "schedule",
        "cron": "0 * * * *",
        "label": "Hourly"
      },
      "starterPrompt": "Check comments across our channels, surface the ones that need a reply, and draft responses for approval.",
      "model": "claude-sonnet-4-6",
      "effort": "medium",
      "tags": [
        "community",
        "social"
      ],
      "proof": {
        "receipts": "comment_watch.py runs hourly with monitoring logs"
      },
      "feature": false,
      "rarity": "common"
    },
    {
      "slug": "code-reviewer-verifier",
      "jobTitle": "Code Reviewer / Verifier",
      "tagline": "Independently reviews another agent's work in a fresh context and rejects what's wrong. The maker can't grade itself.",
      "agent": [
        "dev"
      ],
      "character": "dario",
      "skills": [
        "code-review",
        "verify"
      ],
      "trigger": {
        "type": "event",
        "event": "task-done",
        "label": "On task-done / PR"
      },
      "starterPrompt": "When a task is marked done or a PR opens, review the work in a fresh context against its goal, run the verifier, and reject (with reasons) anything wrong.",
      "model": "claude-opus-4-8",
      "effort": "high",
      "tags": [
        "engineering",
        "review",
        "verification"
      ],
      "proof": {
        "receipts": "task-queue verify system; live /agentic-loop demo",
        "sampleUrl": "https://5dive.com/agentic-loop"
      },
      "feature": true,
      "rarity": "epic"
    },
    {
      "slug": "chief-of-staff",
      "jobTitle": "Chief of Staff / Executive Assistant",
      "tagline": "Pulls status across the whole org, writes one briefing, and flags what's blocked on you.",
      "agent": [
        "marketing"
      ],
      "character": "theo",
      "skills": [
        "5dive-cli",
        "notify-user"
      ],
      "trigger": {
        "type": "schedule",
        "cron": "0 8 * * *",
        "label": "Daily"
      },
      "starterPrompt": "Pull status across the org, write one concise briefing, and push anything blocked on the human to them.",
      "model": "claude-sonnet-4-6",
      "effort": "medium",
      "tags": [
        "ops",
        "exec"
      ],
      "proof": {
        "receipts": "daily briefing + blocker-push"
      },
      "feature": false,
      "rarity": "rare"
    }
  ]
}