Skip to content

Whetstone

by whetstone.cyberelf.link · remote · no sign-in to list tools · v0.8.0 · published by link.cyberelf.whetstone

Verifier-grounded AI promotion gates, disposable report cards, and signed PASS/HOLD/BLOCK receipts.

Changes · 12
CompareRegistry entry

14 tools: 12 change things and 2 only read.

Scanned Oct 21, 2026

2 checks matched something in the tool definitions. No finding is not a safety guarantee. How the checks work.

claude mcp add --transport http tools https://whetstone.cyberelf.link/mcp

Findings

Matches of fixed checks, with the text they matched.

2 info

Tools

Open a tool for its reasons and full description. Descriptions are the publisher's own words, shown as they are.

14 tools

Changes things 12Claude Code: ask
  • about_whetstoneWhat this service is: the tool catalog, the tier boundaries, and where the source lives.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    What this service is: the tool catalog, the tier boundaries, and where the source lives.
  • audit_leakageExact declared-exposure audit over your exam rows: row identity, behavioral fingerprints for graph-DSL expressions, text-similarity review flags, and a clean exam export.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    Exact declared-exposure audit over your exam rows: row identity, behavioral fingerprints for graph-DSL expressions, text-similarity review flags, and a clean exam export. Full example: GET /api/examples key 'leakage'.
  • bank_healthItem-lifecycle diagnostics over your grading history: discriminators, saturated and flaky items, frontier gaps.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    Item-lifecycle diagnostics over your grading history: discriminators, saturated and flaky items, frontier gaps. Full example: GET /api/examples key 'health'.
  • counterexample_huntBounded simulated-annealing search for a graph counterexample inside a DSL predicate class, with an exact certificate when found.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    Bounded simulated-annealing search for a graph counterexample inside a DSL predicate class, with an exact certificate when found. CPU-bounded and strictly rate-limited. Full example: GET /api/examples key 'counterexample'.
  • open_bench_leaderboardTIER 2: list the self-attested public Open Promotion Bench receipts.No findings

    Why it is in this group

    • The name "open_bench_leaderboard" says "open": a word for a change.
    Description, as the server wrote it
    TIER 2: list the self-attested public Open Promotion Bench receipts. Entries contain manifests, verdicts, item-level transitions, and commitments but never task contents or submitted answers.
  • open_bench_startTIER 2: start a one-shot Open Promotion Bench session.1 finding

    Why it is in this group

    • The name "open_bench_start" says "open": a word for a change.

    Findings

    • Info

      The description mentions "credential".

      Matched text

      description: "...tested track rather than a private-bank credential."

      What this check looks for
    Description, as the server wrote it
    TIER 2: start a one-shot Open Promotion Bench session. Returns six fresh virtual-repository scope-integrity tasks. Run a baseline and candidate independently on the same cohort, then submit both answer maps with open_bench_submit. This is an open, procedural, self-attested track rather than a private-bank credential.
  • open_bench_submitTIER 2: grade paired baseline and candidate patches, count gains/regressions/ties, and issue PASS/HOLD/BLOCK.No findings

    Why it is in this group

    • The name "open_bench_submit" says "open": a word for a change.
    Description, as the server wrote it
    TIER 2: grade paired baseline and candidate patches, count gains/regressions/ties, and issue PASS/HOLD/BLOCK. Set publish=true plus attestation=true to append only the safe manifests and sanitized receipt to the public board; tasks and answers are never persisted.
  • promotion_gatePASS, HOLD, or BLOCK from paired per-item results: gains, regressions, exact McNemar p-value, per-domain breakdown.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    PASS, HOLD, or BLOCK from paired per-item results: gains, regressions, exact McNemar p-value, per-domain breakdown. Full example: GET /api/examples key 'gate'.
  • replay_traceTurn reasoning-emulator control events into checkpoints, rewinds, notes, and a timeline.No findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    Description, as the server wrote it
    Turn reasoning-emulator control events into checkpoints, rewinds, notes, and a timeline. Full example: GET /api/examples key 'replay'.
  • report_card_startTIER 1: start a disposable report-card session.1 finding

    Why it is in this group

    • The name "report_card_start" says "start": a word for a change.

    Findings

    • Info

      The description mentions "credential".

      Matched text

      description: "...posable items; it is not a private-bank credential."

      What this check looks for
    Description, as the server wrote it
    TIER 1: start a disposable report-card session. Returns exam items (graph-repair prompts minted from the repository's public frontier) for THIS agent to answer. Answer every item, then call report_card_submit exactly once. Sessions are one-shot, expire in 15 minutes, and are strictly rate-limited. This demonstrates the promotion-gate mechanism on disposable items; it is not a private-bank credential.
  • report_card_submitTIER 1: submit answers for a report-card session and receive the graded report (per-item verdicts, per-domain totals, SHA-256 commitments).No findings

    Why it is in this group

    • The name "report_card_submit" says "submit": a word for a change.
    Description, as the server wrote it
    TIER 1: submit answers for a report-card session and receive the graded report (per-item verdicts, per-domain totals, SHA-256 commitments). Grading is by checker spec: verified strict refinements are reported separately, and promotion grade requires at least 5% clean-support retention. No answer key exists. The session is destroyed by this call.
  • safe_patchApply a section-scoped Markdown patch under conservation checks (untouched sections stay byte-identical; protected tokens preserved).No findings

    Why it is in this group

    • The name "safe_patch" says "patch": a word for a change.
    Description, as the server wrote it
    Apply a section-scoped Markdown patch under conservation checks (untouched sections stay byte-identical; protected tokens preserved). Full example: GET /api/examples key 'safepatch'.
Only reads 2Claude Code: allow
  • inspect_promotionQuarantine declared exposure, compare paired baseline/candidate outcomes on the clean remainder, and issue a promotion receipt.No findings

    Why it is in this group

    • The name "inspect_promotion" says "inspect": a word for reading.
    Description, as the server wrote it
    Quarantine declared exposure, compare paired baseline/candidate outcomes on the clean remainder, and issue a promotion receipt. Bring your own exam rows, exposure records, and per-item results. Full example: GET /api/examples key 'inspector'.
  • memory_relevanceCompare query-free salience against objective-conditioned relevance for a set of memories under a token budget.No findings

    Why it is in this group

    • The description's first verb is "compare": a word for reading.
    Description, as the server wrote it
    Compare query-free salience against objective-conditioned relevance for a set of memories under a token budget. Full example: GET /api/examples key 'memory'.

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Install

Pinned to version 0.8.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http tools https://whetstone.cyberelf.link/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__tools__inspect_promotion",
      "mcp__tools__memory_relevance"
    ],
    "ask": [
      "mcp__tools__audit_leakage",
      "mcp__tools__promotion_gate",
      "mcp__tools__bank_health",
      "mcp__tools__safe_patch",
      "mcp__tools__counterexample_hunt",
      "mcp__tools__replay_trace",
      "mcp__tools__report_card_start",
      "mcp__tools__report_card_submit",
      "mcp__tools__open_bench_start",
      "mcp__tools__open_bench_submit",
      "mcp__tools__open_bench_leaderboard",
      "mcp__tools__about_whetstone"
    ],
    "deny": []
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. It assumes the server is installed as tools. Review it before you use it.