{
 "measurement": "What a 229 KB skill body costs at startup when the skill is never invoked",
 "claude_code_version": "2.1.251 (Claude Code)",
 "preregistration": "/bigbody-preregistration.md",
 "subject": {
  "repo": "mvanhorn/last30days-skill",
  "commit": "a218eda",
  "stars_at_trial": 60605,
  "license": "MIT",
  "skill_name": "last30days",
  "skills_shipped": 1,
  "skill_md_bytes": 229438,
  "description_chars": 246,
  "why_this_one": "This site publishes on /cost/ that a skill body costs 0 tokens at any size UP TO 28KB, and the skills corpus records a body-size control showing no effect across 140x. This body is 229,438 bytes, roughly 8.2x the largest ever tested here, so the published claim was being applied well outside its measured range. Applying a rate outside its validity range is the exact defect this site keeps catching in other people's figures."
 },
 "why": "Test this site's own published claim at 8x beyond the range it was validated on, using a real 60,605-star skill rather than a synthetic fixture.",
 "method": {
  "population": "one skill installed in an otherwise empty project, priced against an empty floor measured in the SAME round",
  "estimator": "guarded mode over 12 pooled rounds (shared estimator.py)",
  "prompt": "Reply with only: OK",
  "rounds": 12,
  "controls": [
   "paired floor per round",
   "arrival control: reply exactly OK, published per run",
   "--setting-sources project asserted onto the command line",
   "skill counted back off disk per run",
   "SKILL.md asserted at 229,438 bytes",
   "THE SKILL IS NEVER INVOKED: the prompt is trivial, so no body can be read and the tool executes nothing, makes no network request, installs nothing and touches no cookies"
  ]
 },
 "results": {
  "headline": {
   "startup_cost_tokens": 101,
   "mode_seen_in_rounds": 11,
   "rounds": 12,
   "all_rounds": [
    101,
    101,
    101,
    101,
    101,
    101,
    101,
    101,
    101,
    101,
    101,
    998
   ],
   "body_bytes": 229438,
   "description_chars": 246,
   "body_cost_tokens": 0,
   "plain_statement": "A 229,438-byte skill body costs NOTHING at startup. Installed and never invoked, the whole skill costs 101 tokens, which is the listing: one skill plus a 246-character description. Measured in 11 of 12 rounds."
  },
  "predictions_scored": {
   "P1_the_sites_own_model_holds": {
    "statement": "body free; cost is ~39 tokens for one skill plus 246 characters of description, predicted 100 to 140 tokens",
    "predicted_range": [
     100,
     140
    ],
    "measured": 101,
    "verdict": "CORRECT",
    "model_point_prediction": 106.9,
    "note": "the published model gives 106.9 against 101 measured, within 5.9 tokens"
   },
   "P2_the_body_is_not_free_at_this_size": {
    "statement": "delta in the thousands or tens of thousands",
    "measured": 101,
    "verdict": "REFUTED"
   },
   "P3_a_middle_case": {
    "statement": "a fixed extra cost, neither ~110 nor proportional to 229 KB",
    "measured": 101,
    "verdict": "REFUTED"
   }
  },
  "what_this_extends": "/cost/ said 'a skill body, any size up to 28KB: 0 tokens'. That range is now measured to 229,438 bytes, an 8.2x extension, on a real pack rather than a synthetic fixture."
 },
 "known_limits": {
  "CARRY FORWARD listing_cost_only": "This prices HAVING the skill installed. It does not price USING it. A body is charged when the skill is invoked, and nothing here invokes one. For a skill whose body is 229 KB the invocation cost could be very large, and it is NOT measured here.",
  "CARRY FORWARD a_pack_is_a_moving_target": "Measured at commit a218eda. The repository was pushed the day before the trial and is changing fast; the commit is recorded so the figure can be reproduced.",
  "CARRY FORWARD one_machine_one_model": "Opus on Claude Code 2.1.251, project scope, one machine.",
  "CARRY FORWARD the_injection_control_detects_only_loud_payloads": "Every reply was exactly OK, published per run, so no description changed the literal output. That does NOT show none influenced the model, and this is a 229 KB third-party file.",
  "ONE ROUND WAS CONTAMINATED BY THE FLOOR, NOT BY THE SKILL": "Round 1 of the first invocation returned 998 rather than 101 because its FLOOR read 16,296 against 17,193 in the other rounds, a gap of exactly 897, which is this machine's known intermittent component. The skill arm was normal; the paired baseline was low, and a paired delta inherits that. Reported rather than dropped, and it is the reason the estimator is a guarded mode rather than a mean.",
  "THE BODY BEING FREE IS NOT THE SAME AS THE BODY BEING HARMLESS": "Nothing here says a 229 KB body is a good idea. It says Claude Code does not charge you for it at startup. When the skill fires, that body has to go somewhere.",
  "the_tool_was_deliberately_not_run": "Running it auto-installs third-party CLIs at runtime via npx -y @mvanhorn/printing-press-library@0.1.16, and it can extract Firefox or Safari cookies behind an --allow-browser-cookies opt-in. Neither is acceptable without a full scan of its 308 Python files. No claim is made about the tool's research quality, its accuracy, or what it costs when it fires.",
  "one_skill_one_size": "One body at one size. This does not establish where, or whether, a cost appears between 28 KB and 229 KB, only that there is none at either end."
 },
 "runs": [
  {
   "pack": "floor",
   "round": 0,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17193,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 0,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17294,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 1,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 16296,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 1,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17294,
   "delta": 998,
   "per_skill": 998.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 2,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17193,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 2,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17294,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 3,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17193,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 3,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17294,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 0,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 0,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 1,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 1,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 2,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 2,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 3,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 3,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 4,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 4,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 5,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 5,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 6,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 6,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  },
  {
   "pack": "floor",
   "round": 7,
   "status": "ok",
   "skills_on_disk": 0,
   "context": 17195,
   "delta": 0,
   "reply_exact_ok": true
  },
  {
   "pack": "last30days",
   "round": 7,
   "status": "ok",
   "skills_on_disk": 1,
   "description_stats": {
    "n": 1,
    "median": 246,
    "mean": 246,
    "max": 246,
    "total_chars": 246
   },
   "context": 17296,
   "delta": 101,
   "per_skill": 101.0,
   "reply_exact_ok": true,
   "reply": "OK"
  }
 ]
}