What makes Coding Pronoun Prompt Resolver notable?

1K downloads. Trust evidence available

How should Coding Pronoun Prompt Resolver be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

Coding Pronoun Prompt Resolver

Detects ambiguous pronouns, vague referents, and bare imperatives in user messages and flags them for resolution using conversation context. Zero-latency det... Skill: Coding Pronoun Prompt Resolver Owner: kaicianflone Summary: Detects ambiguous pronouns, vague referents, and bare imperatives in user messages and flags them for resolution using conversation context. Zero-latency det... Tags: latest:0.11.0 Version history: v0.11.0 | 2026-05-31T16:09:22.001Z | user Always-on logging directive + locked, sanitizing ledger writer (bin/log-resolution.py). Secret/PII redaction, hex

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

1.0k

Updated

Oct 11, 2026

Version

0.11.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1K downloads reported by the source. Last updated 10/11/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 11, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 11, 2026
Adoption signal: 1K downloadsadoption · observed Oct 11, 2026
Latest release: 0.11.0release · observed May 31, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s173yh69sk2g7jzav29e74412s83qaga:pronoun-resolver

Setup complexity is LOW. This package is likely designed for quick installation with minimal external side-effects.
Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-kaicianflone-pronoun-resolver/snapshot"

Run-check

$0.02 USD

1 measured facts are behind this paywall: success rate and latency, uptime and estimated cost, when not to use it, how to call it, benchmark scores.

Agents pay $0.02 in USDC. A card payment is $0.50, the smallest a card allows.

Documentation

CLAWHUB

116,471 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: pronoun-resolver
version: 0.11.0
description: |
  Detects ambiguous pronouns, vague referents, and bare imperatives in user messages
  and flags them for resolution using conversation context. Zero-latency detection via
  hook; resolution happens inside the conversation where context lives. Self-learning
  via correction ledger with adaptive confidence tiering.
capabilities:
  - user-prompt-submit hook (fires on every message, regex-only, no LLM calls)
  - file-write: ~/.claude/skills/pronoun-resolver/.claude/pronoun-ledger.json (resolution metadata, no raw prompts)
  - file-write: ~/.claude/skills/pronoun-resolver/.claude/pronoun-resolver-analytics.jsonl (per-message stats)
data_retention: |
  All data is local-only, never transmitted externally. Prompts are hashed (SHA-256),
  never stored as text. Both data files can be deleted without affecting functionality.
hooks:
  user-prompt-submit:
    - type: command
      command: "bash ${CLAUDE_SKILL_DIR}/bin/detect-pronouns.sh"
      statusMessage: "Scanning for ambiguous references..."
---

# Pronoun Resolver

## Arguments

If invoked with `--stats`: run `python3 ~/.claude/skills/pronoun-resolver/bin/stats.py` and display the output. Do not proceed with the rest of this skill.

---

You are operating with the pronoun resolver active. When the hook detects ambiguous
references in a user message, you will see flags injected before the message.

## Your Role

YOU are the resolver. You have the conversation context. The hook just detects — you decide.

## Resolution Tiering

When you see `[AMBIGUOUS:]` flags, apply this framework:

### GREEN — Resolve silently (90%+ confidence)
The referent is obvious from the last 1-3 messages. Just act. Don't mention the resolution.
- "Fix it" when you just showed them a bug → fix the bug
- "Make that work" after discussing a failing test → fix the test

### YELLOW — State assumption, proceed (70-90% confidence)
You're fairly sure but there's ambiguity. State what you're assuming in one line, then act.
- "I'm taking 'the other one' to mean `auth.ts` since we discussed two files. Acting on that."

### RED — Ask before acting (<70% confidence)
Multiple plausible referents, or no recent context to resolve against. Ask concisely.
- "What should I make good — the UI layout we discussed or the API response format?"

### BLACK — Bare imperative, no context at all
First message of a conversation with no object. Always ask.
- "Make good" with no prior context → "What would you like me to improve?"

## Flag Format

The hook outputs a preamble followed by flags:
```
[PRONOUN-RESOLVER: Resolve these using conversation context. HIGH confidence=act silently. MEDIUM=state assumption then act. LOW/no context=ask user first.]
[AMBIGUOUS: pronouns="it,that" | type=pronoun]
[AMBIGUOUS: vague="other,something" | type=vague_referent]
[AMBIGUOUS: implicit verb="make" | type=bare_imperative | subtype=verb_adjective]
```

## Ledger

Resolution accuracy is tracked at `~/.claude/ski

README.md

# Pronoun Resolver

A Claude Code hook that detects ambiguous references in user prompts and flags them for resolution. Zero-latency detection via regex/heuristics — Claude resolves using its own conversation context.

No LLM calls. No external API keys. Fires on every message, produces output only when ambiguity is detected.

## The Problem

When you type "fix it", Claude has to guess what "it" refers to. Sometimes it guesses right. Sometimes it confidently refactors the wrong file.

When you type "Make good" with no context, Claude may invent an interpretation rather than asking.

This hook makes the ambiguity visible so Claude asks instead of guessing.

## How It Works

```
User types: "fix it"
                |
    +-----------v-----------+
    |  Hook: detect-pronouns |
    |  (regex + heuristic)   |
    |  ~0ms, no LLM calls    |
    +-----------+-----------+
                | ambiguity detected
    +-----------v-----------+
    |  Output: flags +       |
    |  compact preamble      |
    +-----------+-----------+
                |
    +-----------v-----------+
    |  Claude receives:      |
    |  [PRONOUN-RESOLVER: Resolve using context. HIGH=act. LOW=ask.]  |
    |  [AMBIGUOUS: pronouns="it" | type=pronoun]                      |
    |  fix it                |
    +--------------------+---+
                |
    +-----------v-----------+
    |  Claude resolves using |
    |  conversation context  |
    |  (GREEN/YELLOW/RED)    |
    +------------------------+
```

Claude is the resolver. It has the conversation context. The hook just makes ambiguity explicit.

## Detection Categories

### 1. Personal Pronouns (always flagged)

`it`, `them`, `they`, `its`

These are always referential — they can't be determiners.

### 2. Demonstratives (smart filtering)

`this`, `that`, `these`, `those`

Only flagged when used as standalone pronouns, NOT as determiners:

| Prompt | Flagged? | Why |
|--------|----------|-----|
| `fix this` | Yes | "this" is standalone, no object |
| `fix this bug` | No | "this" is a determiner for "bug" |
| `do that and deploy` | Yes | "that" followed by conjunction |
| `update that file` | No | "that" is a determiner for "file" |
| `these tests are failing` | No | "these" is a determiner for "tests" |

### 3. Vague Referents

`other`, `something`, `someone`, `somewhere`, `anything`, `everything`, `stuff`

### 4. Bare Imperatives (implicit subject)

Detected when no pronouns or vague words are found. Catches commands with no explicit object:

| Prompt | Detected | Subtype |
|--------|----------|---------|
| `Fix` | Yes | bare_verb |
| `Make good` | Yes | verb_adjective |
| `Make better/faster` | Yes | verb_adjective |
| `Clean up` | Yes | verb_adjective |
| `Fix the bug` | No | has explicit object |
| `Add tests` | No | has noun object |

## Resolution Tiering

When Claude sees flags, it applies this framework:

| Tier | Confidence | Action |
|------|-----------|--------|
| GREEN | 90%+ | Resolve silently, just act |
| YE

_meta.json

{
  "ownerId": "kn7aakh8bj9gfbh6ah7rtw6pf180g5qr",
  "slug": "pronoun-resolver",
  "version": "0.11.0",
  "publishedAt": 1780243762001
}

CHANGELOG.md

# Changelog

All notable changes to the Pronoun Resolver skill are documented here.

## [0.11.0] - 2026-05-31

### Added
- **Always-present logging directive.** The hook now injects a
  `[PRONOUN-RESOLVER-LOG: ...]` directive on every flagged message, carrying the
  exact `bin/log-resolution.py` command and ledger path. The hook prints the
  directive; Claude runs it to record a resolution. The directive is now in
  context on every fire even when `SKILL.md` isn't loaded (previously the logging
  instruction lived only in the skill body, so the ledger almost never updated).
- **`bin/log-resolution.py`** — a locked, sanitizing ledger writer that owns all
  ledger mutations.
- **`tests/test_log_resolution.py`** — dependency-free unit tests for the
  sanitizer, ledger I/O, validation, and concurrency.

### Security
- **Secret/PII redaction** on every free-text field before it touches disk:
  API keys, tokens (OpenAI/Stripe/GitHub/GitLab/Slack/npm/Google), JWTs, PEM
  private keys, credentials in DB URLs, Bearer/Basic auth, emails, SSNs, phone
  numbers, and long hex/base64 blobs. Control characters (except tab/newline)
  stripped; over-long values truncated.
- **`prompt_hash` validated** as a hex digest (raw text is dropped), and the
  emitted hook directive single-quotes install-derived paths so a checkout path
  containing shell metacharacters can't become executable syntax.

### Fixed
- **Concurrency race:** ledger writes now run under an exclusive `flock` across
  the full read-modify-write, so parallel hook fires no longer drop entries.
- **Corruption safety:** a non-empty ledger that can't be parsed (or isn't the
  expected object shape) is backed up to `*.corrupt` instead of being silently
  overwritten; `confidence` is clamped to `[0,1]` (non-finite → `0.5`) and the
  writer refuses to emit `NaN`/`Infinity` (`allow_nan=False`).

skill-card.md

## Description:

Detects ambiguous pronouns, vague referents, and bare imperatives in user messages and flags them for resolution using conversation context.

This skill is ready for commercial/non-commercial use.

## Publisher:

[kaicianflone](https://clawhub.ai/user/kaicianflone)

### License/Terms of Use:

MIT-0

## Use Case:

Developers and coding-agent users use this skill to make ambiguous prompts visible before an agent guesses the wrong referent. It guides the agent to resolve high-confidence references from conversation context, state assumptions for moderate confidence, or ask before acting when context is weak.

### Deployment Geography for Use:

Global

## Known Risks and Mitigations:

Risk: The hook runs on every submitted prompt and scans prompt text locally.

Mitigation: Install only where this behavior is acceptable, and use the documented project disable sentinel when prompt scanning should be inactive.

Risk: The skill keeps local analytics and a resolution ledger under the skill directory.

Mitigation: Review the stored fields and delete the ledger or analytics files when retained metadata is no longer wanted.

Risk: Stored free-text resolution metadata may still contain novel secret formats or plain-language personal data despite pattern-based redaction.

Mitigation: Avoid logging sensitive referents, review retained metadata periodically, and clear the local files for sensitive projects.

## Reference(s):

- [ClawHub skill page](https://clawhub.ai/kaicianflone/skills/pronoun-resolver)
- [Publisher profile](https://clawhub.ai/user/kaicianflone)
- [README](artifact/README.md)
- [Changelog](artifact/CHANGELOG.md)
- [Evaluation results](artifact/evals/results.json)

## Skill Output:

**Output Type(s):** [Text, Markdown, Shell commands, Configuration, Guidance]

**Output Format:** [Markdown guidance with inline shell commands and local hook output flags]

**Output Parameters:** [1D]

**Other Properties Related to Output:** [Runs locally as a user-prompt-submit hook; stores local analytics and resolution metadata when the agent records a resolution.]

## Skill Version(s):

0.11.0 (source: SKILL.md frontmatter and release evidence)

## Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

Github ReposUpdated 2d agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.