platform-profile.github.json

{
  "_comment": "ENRICHED github platform profile — extends the THIN github entry already present in the installed growth-os references/platform_profiles.json. This is a repo artifact; merge field-by-field into the skill's platform_profiles.json deliberately, do NOT clobber the existing entry. Base fields kept compatible with the reddit/hackernews/existing-github convention; github-specific extension blocks (growth_surfaces, trending_algorithm, breakout_triggers, readme_first_screen_norms) are ADDED and explicitly non-standard — no other platform carries them. Every stat cites RESEARCH-CORPUS.md. Fields marked TODO(evidence) have no in-corpus source and must not be treated as facts.",
  "name": "GitHub",
  "primary_metric": "stars",
  "secondary_metrics": ["forks", "contributors", "watchers", "downloads"],
  "_secondary_metrics_note": "Kept forks/contributors/watchers to stay compatible with the existing entry's [forks, watchers, issues_opened, contributors]; added 'downloads' (package-registry + release-asset, public) as the adoption proxy per OpenSauced. 'clones' deliberately EXCLUDED from the metric set: Traffic API needs push access, 14-day rolling, no backfill — unavailable for advised (non-owned) repos. See RESEARCH-CORPUS.md and automation-spec.md Gate 5 caveat.",
  "content_format": "repository_readme",
  "max_title_length": {
    "repo_description": 350,
    "hn_show_hn_title": 80,
    "_note": "Split per asset because the rubric's launch-title dimension scores an HN/Reddit title (~80 char practical cap), NOT the repo description. 350 approximates GitHub's About-field limit and is TODO(evidence): not established in RESEARCH-CORPUS.md; corpus-measured description length is median 133, range 26-301 chars. Use the corpus range to score, not 350."
  },
  "supports_media": true,
  "supports_markdown": true,
  "tone": "technical_documentation",
  "best_for_stages": ["launch", "growth"],
  "best_for_categories": ["developer_tools", "open_source", "ai_ml"],
  "_best_for_categories_note": "Matches the existing entry exactly (developer_tools/open_source/ai_ml). 'saas_b2b' was dropped — no corpus basis.",
  "audience_traits": [
    "developers_only",
    "readme_is_the_landing_page",
    "will_inspect_source_and_activity",
    "skeptical_of_marketing",
    "reward_falsifiable_concrete_claims",
    "hardware_and_benchmark_literate"
  ],
  "anti_patterns": [
    "explicit_star_us_cta",
    "star_for_star_rings",
    "purchased_or_fake_stars",
    "coordinated_synchronized_starring",
    "superlative_laden_titles",
    "empty_repo_or_no_license",
    "all_marketing_no_code",
    "demo_gif_as_hero"
  ],
  "posting_rules": {
    "method": "requires_real_repository",
    "bot_risk": "low",
    "rate_limit": "no platform limit; growth is HN/Reddit/Trending-driven, one credible launch per repo",
    "warm_up": "seed first ~100 stars via direct outreach before public launch, then feed velocity via HN Show HN, awesome-list PRs, and Dev.to/Reddit cross-posts (star-history.com playbook, RESEARCH-CORPUS.md M6)"
  },
  "optimal_timing": {
    "days": ["tuesday", "wednesday", "thursday"],
    "hours_utc": [12, 13, 14, 15, 16, 17],
    "notes": "This window governs the HN/link LAUNCH POST that drives GitHub, not GitHub itself (GitHub has no 'post time'). Posting the driving HN launch in the 12-17 UTC window is worth ~200 stars over the worst window (arXiv 2511.04453, 138 AI-repo launches; RESEARCH-CORPUS.md M1). Growth is a spike-then-decay curve, not sustained; cadence (Launch Weeks, milestone posts) converts spikes to slopes."
  },
  "growth_surfaces": {
    "_note": "EXTENSION BLOCK — non-standard; no other platform profile carries this. The copy assets that drive stars, mapped to the rubric.",
    "repo_description": "<=133 char median (corpus range 26-301; RESEARCH-CORPUS.md L-desc). The GitHub-search and social-card one-liner. Converting formulas: 'The open-source X alternative'; quantified-outcome stack ('60-90% fewer tokens. Single binary, zero deps'); 'X for Y' recast; category coinage + definite article. Agent/CLI compatibility roll-call rising (10/30, L-desc).",
    "readme_first_screen": "Above-fold anatomy: logo/banner 14/17, one-line bolded value prop 17/17 (universal), motion-or-static-screenshot demo 10/17, badge cluster 15/17 (median ~5-6), nav row 7/17. NOT a demo GIF (0/17). RESEARCH-CORPUS.md L-readme.",
    "launch_title": "The HN/Show HN or link-post title. Formula: '[Name] – one clear sentence'. Modest language only; superlatives correlate with flagging and violate HN guidelines (RESEARCH-CORPUS.md M4).",
    "social_launch_post": "Loud-founder X post or a third-party demo clip; personality accounts outperform org accounts, and a <30s magic-moment clip is itself the distribution (RESEARCH-CORPUS.md C-pattern 3,5).",
    "release_notes": "Milestone/Launch-Week cadence posts that re-trigger star velocity after the initial spike decays (RESEARCH-CORPUS.md C-pattern 6)."
  },
  "trending_algorithm": {
    "_note": "EXTENSION BLOCK — non-standard. GitHub's Trending algorithm is unpublished; this is community analysis (RESEARCH-CORPUS.md M-trending).",
    "basis": "star_velocity_relative_to_repo_baseline",
    "window": "rolling (short; ~24-48h feed from an HN spike)",
    "normalization": "language-normalized (10k stars in 2016 != 10k in 2025; per-language cohort)",
    "note": "Community analysis (github.com/orgs/community/discussions/163970): Trending ranks acceleration off a repo's own baseline, not absolute totals — a small repo with a sharp HN-fed spike can outrank a large steady one. HN->Trending flywheel: HN score is the strongest launch-day predictor (arXiv 2511.04453), baseline stars second."
  },
  "breakout_triggers": {
    "_note": "EXTENSION BLOCK — non-standard. Ranked by cross-case frequency (RESEARCH-CORPUS.md C-patterns). These are DISTRIBUTION mechanics, mostly OUTSIDE copy — the copy rubric only sets them up. The distribution playbook (distribution.md) is where these are worked.",
    "ranked": ["wave_riding", "comparative_number_oneliner", "demo_as_copy", "hn_trending_flywheel", "loud_founder", "cadence"]
  },
  "readme_first_screen_norms": {
    "_note": "EXTENSION BLOCK — non-standard. Measured frequencies, N=17 breakout READMEs, 2026-07-07 (RESEARCH-CORPUS.md L-readme).",
    "one_line_value_prop": "17/17 universal",
    "logo_or_banner_above_fold": "14/17 near-universal",
    "badge_cluster_above_fold": "15/17 expected (median ~5-6)",
    "motion_or_screenshot_demo_above_fold": "10/17 recommended (hosted MP4 or static hero, NOT GIF)",
    "demo_gif_above_fold": "0/17 obsolete",
    "nav_link_row_above_fold": "7/17 optional",
    "comparison_or_why_not_x_section": "9/17 common",
    "benchmark_table": "7/17 common",
    "star_history_chart": "10/17 near-bottom (does the star-asking implicitly)",
    "explicit_star_cta": "1/17 avoid",
    "first_install_command_line": "median line 84.5 (range 27-309, n=16); only 6/16 within first 60 lines — do NOT over-penalize a deferred install"
  }
}