@buoy-gg/agent-core 7.0.40 → 7.0.42

package/lib/commonjs/engine/tokenCalibration.js

@@ -1,81 +1 @@
-"use strict";
-
-Object.defineProperty(exports, "__esModule", {
-  value: true
-});
-exports.TokenCalibration = exports.DEFAULT_CHARS_PER_TOKEN = void 0;
-exports.promptTokensOf = promptTokensOf;
-/**
- * How many characters make a token — measured, not assumed.
- *
- * Every budget in this engine is kept in characters because that is what a
- * string has, but every LIMIT that matters is in tokens: the provider's
- * context window, the reply room, the cost. The conversion was a constant
- * (4) everywhere. English prose does run near 4; the tool results this agent
- * actually sends are JSON, which tokenizes nearer 3 — so a request the budget
- * called 100k tokens could be 130k, on a gateway whose window is 128k. The
- * proactive budget was optimistic in exactly the sessions it exists for.
- *
- * The provider tells us the truth after every request: its usage report
- * carries the prompt's token count, and the engine knows how many characters
- * it sent. One division gives the ratio for THIS conversation on THIS
- * tokenizer; an exponential average smooths a single odd request (a round
- * that was all prose, or all ids). Strands' design 0008 does the same in
- * reverse — a known token baseline plus an estimate of the delta — and its
- * decision record says why: public limits should be in LLM-native units.
- *
- * Clamped, because a wrong sample must not be able to blow the budget wide
- * open (a gateway that reports nonsense usage) or shrink it to nothing.
- */
-
-const DEFAULT_CHARS_PER_TOKEN = exports.DEFAULT_CHARS_PER_TOKEN = 4;
-const MIN_CHARS_PER_TOKEN = 2.5;
-const MAX_CHARS_PER_TOKEN = 5;
-/** Weight of the newest sample once there is history. The first sample replaces the default outright. */
-const EMA_ALPHA = 0.3;
-/** Below this many prompt tokens a sample says more about rounding than about the tokenizer. */
-const MIN_SAMPLE_TOKENS = 500;
-class TokenCalibration {
-  ratio = DEFAULT_CHARS_PER_TOKEN;
-  samples = 0;
-
-  /** Characters per token, as currently believed. */
-  get charsPerToken() {
-    return this.ratio;
-  }
-  get sampleCount() {
-    return this.samples;
-  }
-
-  /** Fold in one request: what was sent, and what the provider counted. */
-  observe(requestChars, promptTokens) {
-    if (!(requestChars > 0) || !(promptTokens >= MIN_SAMPLE_TOKENS)) return;
-    const sample = Math.min(MAX_CHARS_PER_TOKEN, Math.max(MIN_CHARS_PER_TOKEN, requestChars / promptTokens));
-    this.ratio = this.samples === 0 ? sample : this.ratio + EMA_ALPHA * (sample - this.ratio);
-    this.samples += 1;
-  }
-
-  /** Characters a token budget allows, at the believed ratio. */
-  chars(tokens) {
-    return Math.floor(tokens * this.ratio);
-  }
-  reset() {
-    this.ratio = DEFAULT_CHARS_PER_TOKEN;
-    this.samples = 0;
-  }
-}
-
-/**
- * The whole prompt's token count from a usage report.
- *
- * Anthropic reports the cached and freshly-written prefix SEPARATELY from
- * `input`; OpenAI-shaped providers fold `cached_tokens` INTO `prompt_tokens`.
- * See TokenUsage. Adding on the wrong protocol double-counts the cache and
- * halves the believed ratio.
- */
-exports.TokenCalibration = TokenCalibration;
-function promptTokensOf(usage, protocol) {
-  if (protocol === "anthropic") return usage.input + (usage.cacheRead ?? 0) + (usage.cacheWrite ?? 0);
-  return usage.input;
-}
-//# sourceMappingURL=tokenCalibration.js.map
+"use strict";Object.defineProperty(exports,"__esModule",{value:true});exports.TokenCalibration=exports.DEFAULT_CHARS_PER_TOKEN=void 0;exports.promptTokensOf=promptTokensOf;const DEFAULT_CHARS_PER_TOKEN=exports.DEFAULT_CHARS_PER_TOKEN=4;const MIN_CHARS_PER_TOKEN=2.5;const MAX_CHARS_PER_TOKEN=5;const EMA_ALPHA=.3;const MIN_SAMPLE_TOKENS=500;class TokenCalibration{ratio=DEFAULT_CHARS_PER_TOKEN;samples=0;get charsPerToken(){return this.ratio}get sampleCount(){return this.samples}observe(t,r){if(!(t>0)||!(r>=MIN_SAMPLE_TOKENS))return;const s=Math.min(MAX_CHARS_PER_TOKEN,Math.max(MIN_CHARS_PER_TOKEN,t/r));this.ratio=this.samples===0?s:this.ratio+EMA_ALPHA*(s-this.ratio);this.samples+=1}chars(t){return Math.floor(t*this.ratio)}reset(){this.ratio=DEFAULT_CHARS_PER_TOKEN;this.samples=0}}exports.TokenCalibration=TokenCalibration;function promptTokensOf(e,t){if(t==="anthropic")return e.input+(e.cacheRead??0)+(e.cacheWrite??0);return e.input}

package/lib/commonjs/engine/verify.js

@@ -1,300 +1,7 @@
-"use strict";
+"use strict";Object.defineProperty(exports,"__esModule",{value:true});exports.VERIFIERS=void 0;exports.atPath=atPath;exports.isSubset=isSubset;exports.verificationTrailer=verificationTrailer;exports.verifyOutcome=verifyOutcome;var _realNow=require("../realNow");const rec=e=>e&&typeof e==="object"&&!Array.isArray(e)?e:void 0;function sleep(e,i){return new Promise(n=>{if(i?.aborted)return n();const t=(0,_realNow.realSetTimeout)(r,e);function r(){(0,_realNow.realClearTimeout)(t);i?.removeEventListener("abort",r);n()}i?.addEventListener("abort",r,{once:true})})}async function poll(e,i,n,t,r){const s=(0,_realNow.realNow)()+n;for(;;){const a=await e();if(i(a))return{value:a,satisfied:true};if(r?.aborted||(0,_realNow.realNow)()+t>s)return{value:a,satisfied:false};await sleep(t,r)}}function isSubset(e,i){if(e===null||typeof e!=="object"||Array.isArray(e))return JSON.stringify(e)===JSON.stringify(i);const n=rec(i);if(!n)return false;for(const[t,r]of Object.entries(e)){if(!(t in n))return false;if(!isSubset(r,n[t]))return false}return true}function atPath(e,i){const n=i.replace(/^\/+/,"").match(/\[[^\]]*\]|[^.[\]/]+/g)??[];let t=e;for(const r of n){if(t===null||typeof t!=="object")return void 0;if(r.startsWith("[")){const s=r.slice(1,-1).trim();if(!Array.isArray(t))return void 0;const a=s.indexOf("=");if(a<0){t=t[Number(s)]}else{const o=s.slice(0,a).trim();const d=s.slice(a+1).trim().replace(/^["']|["']$/g,"");t=t.find(u=>u&&typeof u==="object"&&String(u[o])===d)}}else{t=t[r]}}return t}const brief=e=>{const i=JSON.stringify(e);return i===void 0?"undefined":i.length>80?`${i.slice(0,77)}\u2026`:i};const storageSetItem=async({params:e,dispatch:i})=>{if(typeof e.key!=="string")return void 0;const n=await i("storage","async.getItem",{key:e.key});const t=typeof n==="string"?n:rec(n)?.value;const r=typeof e.value==="string"?e.value:JSON.stringify(e.value);if(t===r)return{status:"verified",detail:`Read back: "${e.key}" now holds the written value.`};return{status:"failed",detail:`Read back after the write: "${e.key}" holds ${brief(t)}, not what was written. The write did not take.`}};const zustandSetState=async({params:e,dispatch:i,after:n})=>{if(typeof e.storeName!=="string")return void 0;const t=rec(n!==void 0?n:await i("zustand","getStoreState",{storeName:e.storeName}));if(!t||t.found===false)return{status:"failed",detail:`Read back: store "${e.storeName}" could not be read after the write.`};const r=t.currentState;if(typeof e.path==="string"){const o=atPath(r,e.path);if(JSON.stringify(o)===JSON.stringify(e.value))return{status:"verified",detail:`Read back: ${e.storeName}.${e.path} is now ${brief(o)}.`};return{status:"failed",detail:`Read back: ${e.storeName}.${e.path} is ${brief(o)}, not ${brief(e.value)}. The store did not take the write.`}}const s=rec(e.state);if(!s)return void 0;if(isSubset(s,r))return{status:"verified",detail:`Read back: "${e.storeName}" now holds the written fields (${Object.keys(s).join(", ")}).`};const a=Object.keys(s).filter(o=>!isSubset(s[o],rec(r)?.[o]));return{status:"failed",detail:`Read back: "${e.storeName}" does not hold the written value for ${a.join(", ")}. The store did not take the write \u2014 read it before deciding what to do.`}};const querySetQueryData=async({params:e,dispatch:i,after:n})=>{const t=e.queryHash!==void 0?{queryHash:e.queryHash}:e.queryKey!==void 0?{queryKey:e.queryKey}:void 0;if(!t)return void 0;const r=rec(n!==void 0?n:await i("query","getQueryData",t));if(!r||r.found===false)return{status:"failed",detail:"Read back: the query could not be read after the write."};const s=r.data;if(typeof e.path==="string"){const a=atPath(s,e.path);if(JSON.stringify(a)===JSON.stringify(e.value))return{status:"verified",detail:`Read back: the cached ${e.path} is now ${brief(a)}.`};return{status:"failed",detail:`Read back: the cached ${e.path} is ${brief(a)}, not ${brief(e.value)}. The cache did not take the write \u2014 the app may have refetched over it.`}}if(e.data===void 0)return void 0;if(isSubset(e.data,s))return{status:"verified",detail:"Read back: the cache now holds the written data."};return{status:"failed",detail:"Read back: the cache does not hold the written data \u2014 the app may have refetched over it, or the shape differed."}};const navigate=async({params:e,dispatch:i,signal:n})=>{if(typeof e.path!=="string")return void 0;const t=e.path.split("?")[0];const{value:r,satisfied:s}=await poll(async()=>rec(await i("route-events","getCurrentRoute",{})),o=>typeof o?.path==="string"&&(o.path===t||o.path.split("?")[0]===t),2e3,250,n);if(s)return{status:"verified",detail:`The app is now on ${t}.`};const a=typeof r?.path==="string"?r.path:"an unknown route";return{status:"failed",detail:`Two seconds after navigating, the app is on ${a}, not ${t}. Do not navigate again blindly \u2014 check the route exists (the sitemap) and whether something redirected.`}};const upsertOverrideRule=async({params:e,result:i,dispatch:n,signal:t})=>{const r=rec(rec(i)?.rule)?.id??rec(e.rule)?.id;if(typeof r!=="string")return void 0;const s=async()=>{const d=rec(await n("network","listOverrideRules",{}));const u=Array.isArray(d?.rules)?d.rules:[];return u.map(rec).find(f=>f?.id===r)};const{value:a,satisfied:o}=await poll(s,d=>typeof d?.hits==="number"&&d.hits>0,3e3,500,t);if(!a)return{status:"failed",detail:"The rule is not in the override list after the write. It did not install."};if(a.enabled===false)return{status:"failed",detail:"The rule is installed but disabled, so it will not fire."};if(o)return{status:"verified",detail:`The rule has matched ${a.hits} request${a.hits===1?"":"s"} \u2014 the app has already received the overridden response.`};return{status:"unverified",detail:"The rule is installed, but no request has matched it yet \u2014 the app is still showing whatever it loaded before. A refetch or a visit to the screen that fetches this will exercise it; until then, do not say the screen shows the override."}};const VERIFIERS=exports.VERIFIERS={"storage.async.setItem":storageSetItem,"zustand.setState":zustandSetState,"query.setQueryData":querySetQueryData,"route-events.navigate":navigate,"network.upsertOverrideRule":upsertOverrideRule};async function verifyOutcome(e){const i=VERIFIERS[`${e.toolId}.${e.action}`];if(!i)return void 0;try{return await i(e)}catch{return void 0}}function verificationTrailer(e){switch(e.status){case"verified":return`
 
-Object.defineProperty(exports, "__esModule", {
-  value: true
-});
-exports.VERIFIERS = void 0;
-exports.atPath = atPath;
-exports.isSubset = isSubset;
-exports.verificationTrailer = verificationTrailer;
-exports.verifyOutcome = verifyOutcome;
-/**
- * Outcome verification: after a write lands, check that what the user asked
- * for actually happened — not just that the adapter said ok.
- *
- * "Install override → `{ok:true}`" was reported as done. But installed ≠ a
- * request has used it ≠ the cart shows it, and each step can succeed while
- * the outcome fails: the app had the old response cached, the screen kept
- * saying In Stock, and the reply said "the item is now sold out". Strands'
- * GoalLoop frames this as a completion criterion with bounded correction; the
- * criterion borrowed here is DETERMINISTIC — a read-back the engine performs
- * with the same tools the model has — not a second model judging the reply.
- *
- * Three outcomes, each honest about what it is:
- *
- * - `verified`   — the read-back shows the requested state.
- * - `unverified` — the write landed but the read-back cannot yet show the
- *                  outcome (an override is installed but no request has
- *                  matched it). Reported as exactly that, never as proof.
- * - `failed`     — the read-back contradicts the write. The model is told
- *                  what was found and asked for ONE correction through the
- *                  same policy and ledger path; never to repeat the write.
- *
- * Bounded by construction: each verifier makes a few reads, polls for at most
- * `waitMs` (small, per action), and stops on the turn's abort signal. It never
- * dispatches a write, and its reads are not stashed as evidence. Undefined
- * means "no opinion" — nothing is appended, which is what every action
- * without a verifier gets.
- */
+[Buoy] Verified: ${e.detail}`;case"unverified":return`
 
-const rec = v => v && typeof v === "object" && !Array.isArray(v) ? v : undefined;
-function sleep(ms, signal) {
-  return new Promise(resolve => {
-    if (signal?.aborted) return resolve();
-    const t = setTimeout(done, ms);
-    function done() {
-      clearTimeout(t);
-      signal?.removeEventListener("abort", done);
-      resolve();
-    }
-    signal?.addEventListener("abort", done, {
-      once: true
-    });
-  });
-}
+[Buoy] Unverified: ${e.detail}`;case"failed":return`
 
-/** Poll `read` every `everyMs` until `until` says so or `waitMs` is spent. Returns the last read. */
-async function poll(read, until, waitMs, everyMs, signal) {
-  const deadline = Date.now() + waitMs;
-  for (;;) {
-    const value = await read();
-    if (until(value)) return {
-      value,
-      satisfied: true
-    };
-    if (signal?.aborted || Date.now() + everyMs > deadline) return {
-      value,
-      satisfied: false
-    };
-    await sleep(everyMs, signal);
-  }
-}
-
-/** Every key in `want` present in `have` with an equal (deep) value. Arrays compare whole. */
-function isSubset(want, have) {
-  if (want === null || typeof want !== "object" || Array.isArray(want)) return JSON.stringify(want) === JSON.stringify(have);
-  const h = rec(have);
-  if (!h) return false;
-  for (const [k, v] of Object.entries(want)) {
-    if (!(k in h)) return false;
-    if (!isSubset(v, h[k])) return false;
-  }
-  return true;
-}
-
-/**
- * The adapters' path grammar — dotted keys, `[0]`, and the id-addressed
- * `[field=value]` the prompt recommends for list edits ("lines[lineId=seed-1].qty").
- * A mirror of shared/sync/valuePath.ts, kept here because agent-core imports
- * nothing. The first live run wrote exactly that form and this resolver, then
- * dotted-only, reported a landed edit as "did not take".
- */
-function atPath(root, path) {
-  const segs = path.replace(/^\/+/, "").match(/\[[^\]]*\]|[^.[\]/]+/g) ?? [];
-  let cur = root;
-  for (const seg of segs) {
-    if (cur === null || typeof cur !== "object") return undefined;
-    if (seg.startsWith("[")) {
-      const inner = seg.slice(1, -1).trim();
-      if (!Array.isArray(cur)) return undefined;
-      const eq = inner.indexOf("=");
-      if (eq < 0) {
-        cur = cur[Number(inner)];
-      } else {
-        const field = inner.slice(0, eq).trim();
-        const value = inner.slice(eq + 1).trim().replace(/^["']|["']$/g, "");
-        cur = cur.find(el => el && typeof el === "object" && String(el[field]) === value);
-      }
-    } else {
-      cur = cur[seg];
-    }
-  }
-  return cur;
-}
-const brief = v => {
-  const s = JSON.stringify(v);
-  return s === undefined ? "undefined" : s.length > 80 ? `${s.slice(0, 77)}…` : s;
-};
-
-// ── the verifiers ────────────────────────────────────────────────────────────
-
-const storageSetItem = async ({
-  params,
-  dispatch
-}) => {
-  if (typeof params.key !== "string") return undefined;
-  const back = await dispatch("storage", "async.getItem", {
-    key: params.key
-  });
-  const value = typeof back === "string" ? back : rec(back)?.value;
-  const wanted = typeof params.value === "string" ? params.value : JSON.stringify(params.value);
-  if (value === wanted) return {
-    status: "verified",
-    detail: `Read back: "${params.key}" now holds the written value.`
-  };
-  return {
-    status: "failed",
-    detail: `Read back after the write: "${params.key}" holds ${brief(value)}, not what was written. The write did not take.`
-  };
-};
-const zustandSetState = async ({
-  params,
-  dispatch,
-  after
-}) => {
-  if (typeof params.storeName !== "string") return undefined;
-  const back = rec(after !== undefined ? after : await dispatch("zustand", "getStoreState", {
-    storeName: params.storeName
-  }));
-  if (!back || back.found === false) return {
-    status: "failed",
-    detail: `Read back: store "${params.storeName}" could not be read after the write.`
-  };
-  const current = back.currentState;
-  if (typeof params.path === "string") {
-    const got = atPath(current, params.path);
-    if (JSON.stringify(got) === JSON.stringify(params.value)) return {
-      status: "verified",
-      detail: `Read back: ${params.storeName}.${params.path} is now ${brief(got)}.`
-    };
-    return {
-      status: "failed",
-      detail: `Read back: ${params.storeName}.${params.path} is ${brief(got)}, not ${brief(params.value)}. The store did not take the write.`
-    };
-  }
-  const want = rec(params.state);
-  if (!want) return undefined;
-  if (isSubset(want, current)) return {
-    status: "verified",
-    detail: `Read back: "${params.storeName}" now holds the written fields (${Object.keys(want).join(", ")}).`
-  };
-  const wrong = Object.keys(want).filter(k => !isSubset(want[k], rec(current)?.[k]));
-  return {
-    status: "failed",
-    detail: `Read back: "${params.storeName}" does not hold the written value for ${wrong.join(", ")}. The store did not take the write — read it before deciding what to do.`
-  };
-};
-const querySetQueryData = async ({
-  params,
-  dispatch,
-  after
-}) => {
-  const key = params.queryHash !== undefined ? {
-    queryHash: params.queryHash
-  } : params.queryKey !== undefined ? {
-    queryKey: params.queryKey
-  } : undefined;
-  if (!key) return undefined;
-  const back = rec(after !== undefined ? after : await dispatch("query", "getQueryData", key));
-  if (!back || back.found === false) return {
-    status: "failed",
-    detail: "Read back: the query could not be read after the write."
-  };
-  const data = back.data;
-  if (typeof params.path === "string") {
-    const got = atPath(data, params.path);
-    if (JSON.stringify(got) === JSON.stringify(params.value)) return {
-      status: "verified",
-      detail: `Read back: the cached ${params.path} is now ${brief(got)}.`
-    };
-    return {
-      status: "failed",
-      detail: `Read back: the cached ${params.path} is ${brief(got)}, not ${brief(params.value)}. The cache did not take the write — the app may have refetched over it.`
-    };
-  }
-  if (params.data === undefined) return undefined;
-  if (isSubset(params.data, data)) return {
-    status: "verified",
-    detail: "Read back: the cache now holds the written data."
-  };
-  return {
-    status: "failed",
-    detail: "Read back: the cache does not hold the written data — the app may have refetched over it, or the shape differed."
-  };
-};
-const navigate = async ({
-  params,
-  dispatch,
-  signal
-}) => {
-  if (typeof params.path !== "string") return undefined;
-  const wanted = params.path.split("?")[0];
-  const {
-    value,
-    satisfied
-  } = await poll(async () => rec(await dispatch("route-events", "getCurrentRoute", {})), r => typeof r?.path === "string" && (r.path === wanted || r.path.split("?")[0] === wanted), 2_000, 250, signal);
-  if (satisfied) return {
-    status: "verified",
-    detail: `The app is now on ${wanted}.`
-  };
-  const at = typeof value?.path === "string" ? value.path : "an unknown route";
-  return {
-    status: "failed",
-    detail: `Two seconds after navigating, the app is on ${at}, not ${wanted}. Do not navigate again blindly — check the route exists (the sitemap) and whether something redirected.`
-  };
-};
-const upsertOverrideRule = async ({
-  params,
-  result,
-  dispatch,
-  signal
-}) => {
-  const id = rec(rec(result)?.rule)?.id ?? rec(params.rule)?.id;
-  if (typeof id !== "string") return undefined;
-  const find = async () => {
-    const list = rec(await dispatch("network", "listOverrideRules", {}));
-    const rules = Array.isArray(list?.rules) ? list.rules : [];
-    return rules.map(rec).find(r => r?.id === id);
-  };
-  const {
-    value,
-    satisfied
-  } = await poll(find, r => typeof r?.hits === "number" && r.hits > 0, 3_000, 500, signal);
-  if (!value) return {
-    status: "failed",
-    detail: "The rule is not in the override list after the write. It did not install."
-  };
-  if (value.enabled === false) return {
-    status: "failed",
-    detail: "The rule is installed but disabled, so it will not fire."
-  };
-  if (satisfied) return {
-    status: "verified",
-    detail: `The rule has matched ${value.hits} request${value.hits === 1 ? "" : "s"} — the app has already received the overridden response.`
-  };
-  return {
-    status: "unverified",
-    detail: "The rule is installed, but no request has matched it yet — the app is still showing whatever it loaded before. A refetch or a visit to the screen that fetches this will exercise it; until then, do not say the screen shows the override."
-  };
-};
-const VERIFIERS = exports.VERIFIERS = {
-  "storage.async.setItem": storageSetItem,
-  "zustand.setState": zustandSetState,
-  "query.setQueryData": querySetQueryData,
-  "route-events.navigate": navigate,
-  "network.upsertOverrideRule": upsertOverrideRule
-};
-
-/** Run the verifier for this action, if there is one. Never throws: a read that blows up is "no opinion". */
-async function verifyOutcome(input) {
-  const verifier = VERIFIERS[`${input.toolId}.${input.action}`];
-  if (!verifier) return undefined;
-  try {
-    return await verifier(input);
-  } catch {
-    return undefined;
-  }
-}
-
-/** The trailer the model reads. Kept in the `[Buoy] ` trailer form the bench's parser splits on. */
-function verificationTrailer(v) {
-  switch (v.status) {
-    case "verified":
-      return `\n\n[Buoy] Verified: ${v.detail}`;
-    case "unverified":
-      return `\n\n[Buoy] Unverified: ${v.detail}`;
-    case "failed":
-      return `\n\n[Buoy] Check FAILED: ${v.detail} You may correct this ONCE — read the current state first, then make a different, targeted change; never repeat the same write and never claim the outcome happened.`;
-  }
-}
-//# sourceMappingURL=verify.js.map
+[Buoy] Check FAILED: ${e.detail} You may correct this ONCE \u2014 read the current state first, then make a different, targeted change; never repeat the same write and never claim the outcome happened.`}}