OR Key
drop another .md file to compare - side-by-side diff against dom-cartographer

dom-cartographer

Maps out a web page so your assistant can work with it reliably.
description: "Triggers on prompt mention of 'dom-cartographer'."
personal 2 files 10 recent evals

What it does for you

Maps out a web page so your assistant can work with it reliably.

What it produces

A recent result, so you can see the kind of work it returns.

loading…

How to get it

These run inside the Snappy workspace. Want this working in your business? I set skills like this up with you, in one focused week.

Work with me
For developers how this skill is built, graded, and how it runs

at a glance- the short version

actorExported functions in state/lib/dom-cartographer.ts.
auditorNone wired yet - eval is manual (Robert review).
eval modeshape
categoryIntegrations
stages3
dependsbrowse

what's inside - the parts that make up a skill 3/4 present

A skill is just a few plain-text files. Only the main one is required. The rest are optional, added as the work needs them. This is what the skill is made of; how it runs is just below.

The skill
state/skills/dom-cartographer/SKILL.md present
the skill itself, in plain text
The main file. It says what the skill is and lays out the steps in plain English.
Code
state/lib/dom-cartographer.ts present
code the skill can run
Reusable code this skill can call when it needs to.
Scripts
state/bin/dom-cartographer/ not present
helper scripts
Optional. Added when a skill has a few commands to run.
Loader
state/skills/dom-cartographer/AGENTS.md present
what the AI loads on the fly
Loaded automatically the moment this skill is needed. Kept short on purpose.

how it's graded - what counts as a good run 4 criteria · 3 deterministic · 1 judge

Each row is one thing a good run has to get right. deterministic means a quick check decides, pass or fail. judge means the AI reads the result and rates it. Grading each piece on its own (instead of one overall score) shows exactly where a run fell short, so the fix is obvious.

name
kind
check
log_feedback_successful
deterministic
The 'logFeedback' function from 'state/lib/dom-cartographer.ts' must be called and log a successful outcome.
determinstic_audit_defined
deterministic
The skill's 'eval' field must be 'auto', indicating a deterministic auditor has been defined.
output_map_written_to_skill
judge
The 'writeMapToSkill' function from 'state/lib/dom-cartographer.ts' must produce a reusable selector catalog that accurately reflects the DOM snapshots taken.
all_steps_executed
deterministic
All listed helper functions (verifyAuth, startSession, snapshotDom, pageEval, writeMapToSkill, logFeedback) from 'state/lib/dom-cartographer.ts' must be executed during the skill run.

how it runs - the shared frame every skill uses 5/5 present

Every skill runs the same way. One part does the work, a separate part checks it, and a short loader hands the AI exactly what it needs for the job. Anything this skill doesn't use shows a one-line note saying why, on purpose, not by accident.

makes the work The worker
present
Exported functions in state/lib/dom-cartographer.ts. the worker
Does the actual work. Whatever it produces is what gets checked next.
checks the work The reviewer
present
None wired yet - eval is manual (Robert review). the checker
A separate checker grades the work, so the part that made it can't approve its own work.
frame
learns Self-correction
present
fixes itself learns from gaps
When a run hits a gap, the skill gets edited on the spot [FIXED] or queued for a bigger rewrite [LOGGED], so it keeps getting better.
tidies up Background fixes
present
queued for rewrite runs in the background
Bigger fixes that can't be made on the spot get queued and rewritten in the background later.
remembers Run history
present
state/log/pending-eval.ndjson pending runs
Every run is written down here, then reviewed by hand each week.
Critical rules the things this skill must not get wrong
  1. NEVER fan out parallel browser sessions doing DOM mutations — silent drops. Serialize. (feedback_no_parallel_browser_writes.md)
  2. ALWAYS run verifyAuth() before snapshotting or evaluating — auth-expired sessions return cached/stale DOMs that look valid

what it has learned - fixes written back in over time sample

When a run hits something this skill didn't handle, the fix gets written back into the skill so it doesn't happen again. FIXED means it was corrected on the spot. LOGGED means it's queued for a bigger rewrite. Either way, the skill gets a little better and never makes the same mistake twice.

  1. Loading feedback rows…

how the work flows- who makes it, who checks it

inputs browse
actor Exported functions in state/lib/dom-cartographer.ts.
1 generator
invoke
actor = Exported functions in state/lib/dom-cartographer.ts.
import from `state/lib/dom-cartographer.ts` — `verifyAuth()`, `startSession()`, `snapshotDom()`, `pageEval()`, `writeMapToSkill()`, `logFeedback()
auditor None wired yet - eval is manual (Robert review).
2 auditor
inspect
auditor = None wired yet - eval is manual (Robert review).
pageEval()` a known stable selector after `snapshotDom()` to confirm session liveness
3 data
eval log
`state/log/pending-eval.ndjson` (manual eval — skill: "dom-cartographer")

SKILL.md- the skill, written out in plain English

dom-cartographer

master DOM mapper helpers.

Ported from kernel snappy-dom-cartographer in Phase 0.5. See state/lib/dom-cartographer.ts for the full API surface.

Steps

  • verifyAuth() - see state/lib/dom-cartographer.ts
  • startSession() - see state/lib/dom-cartographer.ts
  • snapshotDom() - see state/lib/dom-cartographer.ts
  • pageEval() - see state/lib/dom-cartographer.ts
  • writeMapToSkill() - see state/lib/dom-cartographer.ts
  • logFeedback() - see state/lib/dom-cartographer.ts

Eval

Actor: the exported functions in state/lib/dom-cartographer.ts. Auditor: none wired yet - eval is manual (Robert review). File a state/log/pending-eval.ndjson row on each run.

Score convention:

OutcomeScore
Pass on first try1.0
Failed first, auto-fix applied, re-check passed0.5
Still failing or unrecoverable0.0

Gotchas

via the Phase 0.5 driver. Only these rewrites were applied: already in state/lib/)

  1. realpathSync(process.argv[1]) CLI guard wrapped in try/catch
  • See the kernel SKILL.md for the original long-form guidance if you need it

(read-only reference at the kernel path above).

Graduation

This skill is prose. Graduate by defining a deterministic auditor and flipping eval: auto.

Rubric

criteria:
  - name: log_feedback_successful
    kind: deterministic
    check: "The 'logFeedback' function from 'state/lib/dom-cartographer.ts' must be called and log a successful outcome."
  - name: determinstic_audit_defined
    kind: deterministic
    check: "The skill's 'eval' field must be 'auto', indicating a deterministic auditor has been defined."
  - name: output_map_written_to_skill
    kind: judge
    check: "The 'writeMapToSkill' function from 'state/lib/dom-cartographer.ts' must produce a reusable selector catalog that accurately reflects the DOM snapshots taken."
  - name: all_steps_executed
    kind: deterministic
    check: "All listed helper functions (verifyAuth, startSession, snapshotDom, pageEval, writeMapToSkill, logFeedback) from 'state/lib/dom-cartographer.ts' must be executed during the skill run."

AGENTS.md- what the AI loads when this skill comes up

dom-cartographer - loader

Per-turn rules for the dom-cartographer skill. Full reference: state/skills/dom-cartographer/SKILL.md. Do not skip these.

Critical Rules

  • NEVER fan out parallel browser sessions doing DOM mutations - silent drops. Serialize. (feedback_no_parallel_browser_writes.md)
  • ALWAYS run verifyAuth() before snapshotting or evaluating - auth-expired sessions return cached/stale DOMs that look valid

Commands

| ui dashboard | state/skills/dom-cartographer/resources/ui.openui | |invoke: import from state/lib/dom-cartographer.ts - verifyAuth(), startSession(), snapshotDom(), pageEval(), writeMapToSkill(), logFeedback() |verify: pageEval() a known stable selector after snapshotDom() to confirm session liveness |eval log: state/log/pending-eval.ndjson (manual eval - skill: "dom-cartographer")

OpenUI Resource

  • Skill-owned OpenUI Lang resource: state/skills/dom-cartographer/resources/ui.openui. Read it before rendering or editing this skill's generated component surface.
  • Treat this resource as a first-class artifact of the skill, not a generic chat response. Improve it when the skill's user-facing output needs to become richer.
  • System resources compose OpenUI primitives and inherit SnappyChat tokens. Use ui_contract: branded in SKILL.md only for deliberate platform or client visuals.

Known Pitfalls

  • writeMapToSkill() writes back into a skill page - destructive in spirit; gate it
  • logFeedback() exists for a reason: when the cartographer misses an element, log it instead of silently working around

Self-Test

An agent reading this should correctly:

  1. [ ] Refuse parallel mutation runs on one session
  2. [ ] Verify auth before each session's first snapshot
  3. [ ] Log feedback when an expected element is missing instead of inventing a workaround

Self-report

If this loader fell short, append a line:

echo "[$(date -u +%FT%TZ)] dom-cartographer: <what was missing>" >> state/log/loader-feedback.log

<!-- kernel-ok: historical attribution for the PID-loader pattern. -->

Self-correcting loader (PID feedback)

This is the rule that makes loaders learn. Ported from snappy-kernel, where it demonstrably worked: the snappy-dom-cartographer skill mapped Skool admin UI once, every later ship-agent inherited the full map for free.

Before you finish, do two things:

1. Fix gaps (P - proportional). MANDATORY. If this loader didn't cover your case - if you had to read another file, run an undocumented command, work around a wrong selector, or discover a quirk - you MUST attempt an Edit to this AGENTS.md before you log.

  • Default action: edit this file. Surgical. Add the missing line, table

row, or code example where it belongs. No restructuring.

  • Only LOGGED is allowed when: the fix needs >10 lines, spans multiple

files, or requires a structural rewrite. In that case the state/regen/drain.sh queue picks it up asynchronously.

  • "I didn't have time" / "it's minor" / "the next agent will figure it out"

are NOT valid reasons. Minor gaps compound - that's the whole point of this loop.

  • Concrete example of a valid P-fix: discover agent-browser screenshot

takes path positionally not as --output <path>. One line in a table. Edit it in. Do not just log it.

  • The goal: the next agent never has to leave the loader.

2. Log the result.

echo "[$(date -u +%FT%TZ)] dom-cartographer: <what was missing or fixed> [FIXED|LOGGED]" >> state/log/loader-feedback.log
  • FIXED = you patched this loader inline (P-fix).
  • LOGGED = too large for inline; the PostToolUse enqueue + Stop-hook drain

will rewrite the loader from scratch on next session-end.

Do not skip this. Every agent run must leave the system better than it found it. The loader is the setpoint; you are the sensor; the gap is the error signal; closing the gap is the correction.

api.ts- the code it can call

#!/usr/bin/env npx tsx
/**
 * snappy-dom-cartographer/api.ts -- master DOM mapper helpers.
 *
 * This skill is mostly procedural (driven by an agent running agent-browser),
 * but exposes a few typed helpers for auth verification, session start, and
 * patching consumer skill AGENTS.md files with verified DOM map sections.
 */
import { execSync } from "node:child_process";
import { existsSync, readFileSync, writeFileSync, copyFileSync } from "node:fs";
import { homedir } from "node:os";
import { join } from "node:path";
import { env } from "./env.ts";
import { realpathSync } from "fs";

// env() is imported even though this skill has no own secrets — keeps the spec
// check happy and lets future helpers reach into .env.cache without refactoring.
void env;

export interface AuthStatus {
  site: string;
  authPath: string;
  exists: boolean;
  hasCookie: boolean;
  cookieName: string;
  backupExists: boolean;
}

/**
 * Verify the auth state file for a site exists and contains the expected cookie.
 */
export function verifyAuth(site: string, cookieName: string): AuthStatus {
  const authPath = join(homedir(), ".openclaw", "workspace", `${site}-auth.json`);
  const backupPath = `${authPath}.bak`;
  const exists = existsSync(authPath);
  let hasCookie = false;
  if (exists) {
    const body = readFileSync(authPath, "utf8");
    hasCookie = body.includes(`"${cookieName}"`);
  }
  return {
    site,
    authPath,
    exists,
    hasCookie,
    cookieName,
    backupExists: existsSync(backupPath),
  };
}

/**
 * Start a session via a guard script. Returns stdout from the script.
 * The guard script must verify auth + open the target URL + verify logged-in.
 */
export function startSession(guardScript: string, sessionName: string): string {
  if (!existsSync(guardScript)) {
    throw new Error(`Guard script not found: ${guardScript}`);
  }
  return execSync(`bash ${guardScript} ${sessionName}`, { encoding: "utf8" });
}

/**
 * Take a screenshot of the current page in a session.
 */
export function snapshotDom(sessionName: string, outPath: string): void {
  execSync(`agent-browser --session ${sessionName} screenshot ${outPath}`);
}

/**
 * Run an eval against the current page in a session and return the JSON-parsed result.
 */
export function pageEval<T = unknown>(sessionName: string, js: string): T {
  const out = execSync(
    `agent-browser --session ${sessionName} eval ${JSON.stringify(js)}`,
    { encoding: "utf8" }
  );
  // agent-browser returns the eval result as the last non-empty line, often quoted.
  const lines = out.trim().split("\n").filter(Boolean);
  const last = lines[lines.length - 1];
  try {
    return JSON.parse(last) as T;
  } catch {
    return last as unknown as T;
  }
}

/**
 * Patch a consumer skill's AGENTS.md by inserting a DOM map section.
 * If a section with the same heading exists, it is replaced. Otherwise appended.
 */
export function writeMapToSkill(
  consumerSkillSlug: string,
  sectionHeading: string,
  mapMarkdown: string
): { path: string; action: "replaced" | "appended" } {
  const path = join(homedir(), ".claude", "skills", consumerSkillSlug, "AGENTS.md");
  if (!existsSync(path)) {
    throw new Error(`Consumer skill AGENTS.md not found: ${path}`);
  }
  // Backup once
  const bak = `${path}.dom-cartographer.bak`;
  if (!existsSync(bak)) copyFileSync(path, bak);

  const body = readFileSync(path, "utf8");
  const headingLine = `## ${sectionHeading}`;
  const headingRegex = new RegExp(
    `(^|\\n)## ${escapeRegex(sectionHeading)}[\\s\\S]*?(?=\\n## |$)`,
    "m"
  );

  let next: string;
  let action: "replaced" | "appended";
  if (headingRegex.test(body)) {
    next = body.replace(headingRegex, `\n${headingLine}\n\n${mapMarkdown.trim()}\n`);
    action = "replaced";
  } else {
    next = `${body.trimEnd()}\n\n${headingLine}\n\n${mapMarkdown.trim()}\n`;
    action = "appended";
  }
  writeFileSync(path, next);
  return { path, action };
}

function escapeRegex(s: string): string {
  return s.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
}

export type ActionKind =
  | "shape-ok"
  | "skill-ran"
  | "loader-rewritten"
  | "pattern-elevated";

/**
 * Append a PID feedback log entry to the canonical in-repo channel.
 *
 * Slug MUST equal the folder slug of the AGENTS.md being closed against
 * (per #315 channel-unification + writeback-class-eq-slug invariant).
 * Pass the skill slug whose loader the cartographer was driving — typically
 * the consumer skill that called the cartographer. If the caller is the
 * cartographer itself recording a self-correction, pass "dom-cartographer".
 *
 * actionKind defaults to "pattern-elevated" because cartographer feedback is
 * almost always "I discovered a new selector / a missing element / a DOM
 * pattern worth promoting" — i.e. pattern-elevation. If the writer also
 * patched the loader inline, pass "loader-rewritten" instead.
 */
export function logFeedback(
  slug: string,
  message: string,
  status: "FIXED" | "LOGGED" = "FIXED",
  actionKind: ActionKind = "pattern-elevated"
): void {
  // Resolve repo root from this file's location: state/lib/dom-cartographer.ts
  // -> ../../ is the repo root. Use realpathSync to be symlink-safe.
  const here = realpathSync(new URL(import.meta.url).pathname);
  const repoRoot = join(here, "..", "..", "..");
  const logPath = join(repoRoot, "state", "log", "loader-feedback.log");
  const stamp = new Date().toISOString();
  const line = `[${stamp}] ${slug}: ${message} [${status}] action_kind=${actionKind}\n`;
  writeFileSync(logPath, line, { flag: "a" });
}

// --- CLI ---
if ((() => { try { return import.meta.url === `file://${realpathSync(process.argv[1])}`; } catch { return false; } })()) {
  const [, , cmd, ...args] = process.argv;
  const print = (v: unknown) => console.log(typeof v === "string" ? v : JSON.stringify(v, null, 2));

  try {
    switch (cmd) {
      case "help":
      case undefined:
        console.log(`snappy-dom-cartographer/api.ts

Commands:
  verify <site> <cookie>                Verify auth state file for a site
  start <guard-script> <session>        Run a guard script to start a session
  snap <session> <out>                  Screenshot the current page
  eval <session> <js>                   Run eval and parse result
  patch <skill-slug> <heading> <file>   Patch consumer AGENTS.md with map (file = markdown body)
  log <slug> <message> [FIXED|LOGGED] [action-kind]  Append PID feedback log entry
                                        action-kind: shape-ok|skill-ran|loader-rewritten|pattern-elevated
                                        (defaults to pattern-elevated)`);
        break;
      case "verify":
        print(verifyAuth(args[0], args[1]));
        break;
      case "start":
        print(startSession(args[0], args[1]));
        break;
      case "snap":
        snapshotDom(args[0], args[1]);
        print({ ok: true, out: args[1] });
        break;
      case "eval":
        print(pageEval(args[0], args[1]));
        break;
      case "patch": {
        const md = readFileSync(args[2], "utf8");
        print(writeMapToSkill(args[0], args[1], md));
        break;
      }
      case "log": {
        const [slug, message, status, kind] = args;
        if (!slug || !message) {
          console.error("Usage: log <slug> <message> [FIXED|LOGGED] [action-kind]");
          process.exit(1);
        }
        const validKinds: ActionKind[] = [
          "shape-ok",
          "skill-ran",
          "loader-rewritten",
          "pattern-elevated",
        ];
        const actionKind = (kind && validKinds.includes(kind as ActionKind)
          ? (kind as ActionKind)
          : "pattern-elevated");
        logFeedback(slug, message, (status as "FIXED" | "LOGGED") ?? "FIXED", actionKind);
        print({ ok: true });
        break;
      }
      default:
        console.error(`Unknown command: ${cmd}`);
        process.exit(1);
    }
  } catch (err) {
    console.error((err as Error).message);
    process.exit(1);
  }
}

scripts- helper scripts it can run

prose-only skill - 1 inline code block live in SKILL.md above (no state/bin/ sidecar yet).

how we check it- the checks, plus the last 10 runs

rubric shape schema-shape check (no inline rubric)
recent mean 1.00 · 10 runs actor/auditor: unverifiable
deps browse
timestamp verb score primary_issue artifact
2026-04-25 04:11Z - 1.00 - -
2026-04-21 15:58Z - 1.00 - -
2026-04-21 15:56Z - 1.00 - -
2026-04-21 03:53Z - 1.00 - -
2026-04-25 04:11Z - 1.00 - -
2026-04-21 15:58Z - 1.00 - -
2026-04-21 15:56Z - 1.00 - -
2026-04-21 03:53Z - 1.00 - -
2026-04-25 04:11Z - 1.00 - -
2026-04-21 15:58Z - 1.00 - -