Agent skill

Tau Tool Verification Status

by dpc in dpc/tau

Verify Tau status state and title transitions, validation, acknowledgement persistence, and activation steering around tools and watched agents.

MPL-2.0Auto-check passed

Install Tau Tool Verification Status

skills CLI
$ npx skills add dpc/tau --skill tau-tool-verification-status -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install dpc/tau tau-tool-verification-status --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/dpc/tau.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/tau-tool-verification-status .claude/skills/tau-tool-verification-status && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
tau-tool-verification-status
GitHub stars
105
Token cost
~1.9k tokens
SKILL.md length
1,018 words
Files
1
Skills in repo
15
Repo updated
First seen
Licence
MPL-2.0

At a glance

Verify Tau status state and title transitions, validation, acknowledgement persistence, and activation steering around tools and watched agents.

  • Works in 6 steps: working, a second working title,… → invalid state and omitted title, which… → empty, whitespace-only, and… → …
  • SKILL.md covers Durable expectations, Efficient live plan, Current quirks to check, not… and Report
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Tau Tool Verification Status is an agent skill from dpc/tau. Verify Tau status state and title transitions, validation, acknowledgement persistence, and activation steering around tools and watched agents.

Its SKILL.md is about 1.9k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

The repository describes itself as: Tau Coding Agent - like Pi, but twice as much. The licence is MPL-2.0.

Example prompts

  • “/tau-tool-verification-status”

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

  1. working, a second working title, waiting, recovery to working,
  2. invalid state and omitted title, which may be rejected by tool-schema
  3. empty, whitespace-only, and surrounding-whitespace titles; verify that
  4. LF, CR, tab, another ASCII control, U+2028, and U+2029 in a title;
  5. titles of exactly 160 and 161 UTF-8 bytes after trimming;
  6. a multibyte boundary case to confirm that the limit counts UTF-8 bytes, not

What it can do on your machine

Read from SKILL.md and the folder at commit d2e1955. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Tau Tool Verification Status loads about 1.9k tokens when it runs. Until then it costs about 43 tokens; SKILL.md has 1,018 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~43
When it runs · the whole SKILL.md, loaded when a task matches
~1.9k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from dpc/tau at commit d2e1955, republished under its MPL-2.0 licence (© dpc). 1,018 words, ~1,884 tokens.

Download SKILL.mdSave it as .claude/skills/tau-tool-verification-status/SKILL.md (or your agent's skills folder).
name
tau-tool-verification-status
description
Verify Tau status state and title transitions, validation, acknowledgement persistence, and activation steering around tools and watched agents.
advertise
false

Tau Tool Verification: Status

Load tau-tool-verification first for shared verification and reporting rules. Use this focused plan for the status tool. Pair it with tau-tool-verification-agent-coordination when testing watched agents or direct messages.

Keep durable expectations below separate from observed quirks. Run live probes through the provider and harness; do not infer activation behavior from source or from receipt of a queued notification alone. Give every concurrent probe a distinct nonce.

Durable expectations

  • status accepts working, waiting, blocked, and done. waiting means progress is paused pending an expected self-resolving event; blocked means progress needs external intervention. It trims the title, then requires a non-empty, single-line canonical title of at most 160 UTF-8 bytes with no control characters, U+2028, or U+2029.
  • A successful call clearly reports the accepted state and title. A rejected call must not imply that the requested state became current.
  • Ordinary transitions work in both directions: initial Working, Working title update, Waiting, recovery from Waiting to Working, Blocked, recovery from Blocked to Working, and Done. Watched status projection preserves Waiting as distinct from Blocked.
  • One accepted Working acknowledgement survives routine tool-result rounds. Routine results must not repeatedly steer the agent to acknowledge the same work.
  • Watched-child progress and status updates remain deliverable but do not steer the watcher into status(working).
  • Genuine activating inputs still activate appropriately: a visible user prompt, a direct agent message, and a watched child's final response. Each activation should cause at most one acknowledgement request for that work, not a reminder loop.
  • Steering and watched progress use concise, generic shapes: Reminder: when working on a task use \status` tool to acknowledge it., Your `status` is set to `working` on "<title>". Set it to `done`, `waiting`, or `blocked` to finish or call `wait` when waiting for external events., and Watched agent <agent-id> status: <state> on <title>`.

Efficient live plan

Use several short-lived delegates when possible so event streams and status state do not contaminate one another. Tell delegates not to change the repository and not to comply with a status reminder that would invalidate the probe.

Transitions and validation

Use a nonce-tagged title and record exact results for:

  1. working, a second working title, waiting, recovery to working, blocked, another recovery to working, and done; verify watched projections preserve both paused phases exactly;
  2. invalid state and omitted title, which may be rejected by tool-schema validation before invocation;
  3. empty, whitespace-only, and surrounding-whitespace titles; verify that accepted output uses the trimmed canonical title;
  4. LF, CR, tab, another ASCII control, U+2028, and U+2029 in a title;
  5. titles of exactly 160 and 161 UTF-8 bytes after trimming;
  6. a multibyte boundary case to confirm that the limit counts UTF-8 bytes, not characters.

Do not claim rejected calls preserved or changed state unless a separate observable proves it; the tool has no status-query operation. Flag an error that echoes the requested state and title as ambiguous if readers could mistake them for current state.

Working persistence across routine rounds

Call status(working) once, then perform at least five ordinary tool rounds with no genuine activating input. Mix quick read-only calls rather than using a single batch. Record every acknowledgement directive and status call. Finish with one done transition, remain alive only long enough to detect a bounded post-Done repeat (for example, two further turns), then stop the delegate.

Pass when no routine round requests or causes another Working acknowledgement. Report post-Done directives separately. With no new substantive tool call, there must be no spontaneous reminder. A post-Done self_info, skill, shell, or other substantive tool admission correctly requests one fresh Working acknowledgement; status and wait are exempt. A repeated reminder without another substantive admission is a defect.

Show full SKILL.md (410 more words)Show less
Watched progress isolation

Start one auto-watched child and instruct it to emit three nonce-tagged Working titles separated by harmless tool rounds, then Done and one final response. Correlate each received event with its nonce and classify it as:

  • watched status/progress;
  • watched final response;
  • direct message;
  • status acknowledgement directive.

Do not infer causality merely because two events are adjacent. Strong evidence is an acknowledgement directive source-linked to the watched event in a trace, or a controlled comparison where the directive appears only after that event. Verify that all progress arrives without watcher status steering, while the final response remains a distinct activating input. Require the same generic status: <state> on <title> shape for initial reports and title updates; do not expect sequencing language such as “started” or “updated”.

Genuine activation paths

Probe each path independently with a nonce. Begin with the receiving agent settled in done and no outstanding tools or watched events. For each path, require either an activation-source trace naming that input or exactly one provider turn that emits the expected nonce-bearing acknowledgement/status effect. Observe at least two further bounded turns or events and require zero duplicate activations.

  1. Send a direct parent-to-child agent message that requests one exact nonce-bearing acknowledgement.
  2. Have one watched child produce a nonce-bearing final response separately from its progress; require the watcher’s resulting provider turn to acknowledge that nonce once.
  3. Ask the external user or controlling session to send a nonce-bearing prompt that requests one exact acknowledgement. If no external sender is available, mark this path unavailable rather than simulating or inferring it.

Receipt proves delivery, not necessarily activation classification. Prefer event traces that name the activation source. Record duplicate deliveries, repeated acknowledgement directives, provider retries, and extra turns.

Current quirks to check, not product contracts

These observations can regress or disappear; verify them rather than assuming them:

  • Implementation-level empty and overlong-title errors may echo the requested state and title, which looks like current-state output even though no acceptance occurred.
  • A delegate that reports Done can receive repeated directives to return to Working after subsequent parent/delegate messages or lifecycle events. Bound the observation window and report the exact triggering sequence; do not keep a delegate alive merely to count an unbounded repeat loop.

Report

Provide a compact causal transcript with exact nonces and tool/error wording. Separate:

  • passed durable expectations;
  • failures with the strongest available causal evidence;
  • inconclusive or unavailable activation paths;
  • current quirks, confusing wording, duplicate prompts or notifications, and wasted calls/turns.

© dpc, MPL-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/tau-tool-verification-status of dpc/tau.

Open the folder on GitHubat commit d2e1955

Compare with similar skills

Tau Tool Verification Status next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Tau Tool Verification Status compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Tau Tool Verification Status this skilldpc/tau105—~1.9kAutomated safety check: PassMPL-2.0
View Transitionsthedaviddias/Front-End-Checklist74k—~528Automated safety check: PassMIT
Video Template Frame Glitch Titlenexu-io/open-design100k—~325Automated safety check: PassApache-2.0
Policy Acknowledgementsickn33/agentic-awesome-skills47k1 repos~3.4kAutomated safety check: PassMIT
Meta Titlethedaviddias/Front-End-Checklist74k—~494Automated safety check: PassMIT
Frame Titlethedaviddias/Front-End-Checklist74k—~438Automated safety check: PassMIT

Similar skills

  • View Transitions

    thedaviddias/Front-End-Checklist

    A skill your agent uses when adding page transition animations, image expand/contract effects, shared-element transitions, or improving navigation UX in a single-page or multi-page application.

    74k GitHub stars~528 tokensUpdated 3 days ago
    Auto-check passed
  • Use this plugin when the user wants a "Glitch Title Frame" HyperFrames motion video — Digital glitch, chromatic offset, and data-corruption title frame for video transitions or cyberpunk heroes.

    100k GitHub stars~325 tokensUpdated today
    Media & CreativeAuto-check passed
  • Policy Acknowledgement

    sickn33/agentic-awesome-skills

    Policy acknowledgement register: employee, policy and version, sent and due dates, acknowledged date and flag, days overdue, reminder sent and status.

    47k GitHub starsUsed in 1 repo~3.4k tokens
    Auto-check passed
  • Meta Title

    thedaviddias/Front-End-Checklist

    A skill your agent uses when auditing page metadata or generating SEO-optimized titles.

    74k GitHub stars~494 tokensUpdated 3 days ago
    Frontend & DesignAuto-check passed
  • Frame Title

    thedaviddias/Front-End-Checklist

    A skill your agent uses when reviewing rendered HTML, interactive components, or design-system patterns related to Provide titles for iframes and frames.

    74k GitHub stars~438 tokensUpdated 3 days ago
    Frontend & DesignAuto-check passed
  • Vercel React View Transitions

    sickn33/agentic-awesome-skills

    Guide React and Next.js view transitions, shared element animations, route transitions, transition types, and reduced-motion-safe UI state animation.

    47k GitHub starsUsed in 1 repo~3.1k tokens
    Auto-check passed
  • A skill your agent uses when selfci, Nix CI, coverage, cargo-crap, CRAP-score, crapAbsolute, or crapReport checks fail in Tau, or before changing the cargo-crap gates, thresholds, or flagged complex…

    105 GitHub stars~1.2k tokensUpdated 3 days ago
    Auto-check passed
  • A skill your agent uses when asked to "triage papercuts", review clanker-reported problems, or analyze and clear tau dev papercut reports.

    105 GitHub stars~2.4k tokensUpdated 3 days ago
    Auto-check passed
  • A skill your agent uses when asked to verify Tau harness tools or tool output behavior, especially read, edit, shell/shellcommand, line-oriented output, truncation, metadata headers, UTF-8 handling…

    105 GitHub stars~4.8k tokensUpdated 3 days ago
    Auto-check passed
  • A skill your agent uses when verifying Tau file and command tools: read, edit, replace, applypatch, shell, or shellcommand, including ranges, UTF-8, truncation, diffs, timeouts, mutation safety, and…

    105 GitHub stars~4k tokensUpdated 3 days ago
    Auto-check passed
  • A skill your agent uses when changing or reviewing Tau's static site under site/ and needing visual verification of layout, spacing, colors, alignment, desktop rendering, mobile rendering, or…

    105 GitHub stars~308 tokensUpdated 3 days ago
    Auto-check passed
  • A skill your agent uses when tracing or auditing Tau agent execution, including provider and cache cost, tool/background/wait latency, outer turns, compaction, delegated workflows, or performance…

    105 GitHub stars~658 tokensUpdated 3 days ago
    Auto-check passed

Questions about Tau Tool Verification Status

What does Tau Tool Verification Status do?

Verify Tau status state and title transitions, validation, acknowledgement persistence, and activation steering around tools and watched agents. Tau Tool Verification Status is an agent skill from dpc/tau. Verify Tau status state and title transitions, validation, acknowledgement persistence, and activation steering around tools and watched agents.

How do I install Tau Tool Verification Status in Claude Code?

Run `npx skills add dpc/tau --skill tau-tool-verification-status -a claude-code`. Or copy the skill folder (.agents/skills/tau-tool-verification-status in dpc/tau) into .claude/skills/tau-tool-verification-status in your project. Claude Code loads it when a task matches its description.

How do I install Tau Tool Verification Status in Codex?

Run `npx skills add dpc/tau --skill tau-tool-verification-status -a codex`. Or copy the skill folder (.agents/skills/tau-tool-verification-status in dpc/tau) into .agents/skills/tau-tool-verification-status in your project. Codex loads it when a task matches its description.

Can I use Tau Tool Verification Status in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add dpc/tau --skill tau-tool-verification-status -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/tau-tool-verification-status, .gemini/skills/tau-tool-verification-status, .github/skills/tau-tool-verification-status and .opencode/skills/tau-tool-verification-status in your project.

What does Tau Tool Verification Status need to run?

SKILL.md names no scripts, command-line tools or credentials: Tau Tool Verification Status is instructions for the agent only.

Does Tau Tool Verification Status access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Tau Tool Verification Status safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Tau Tool Verification Status use?

Tau Tool Verification Status is published under the MPL-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Tau Tool Verification Status use?

About 1.9k tokens (SKILL.md is roughly 7.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Tau Tool Verification Status?

Skills that share tags, products or a category with Tau Tool Verification Status: View Transitions (thedaviddias/Front-End-Checklist, 74k stars), Video Template Frame Glitch Title (nexu-io/open-design, 100k stars), Policy Acknowledgement (sickn33/agentic-awesome-skills, 47k stars) and Meta Title (thedaviddias/Front-End-Checklist, 74k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Tau Tool Verification Status?

dpc (a GitHub user) maintains it in dpc/tau, which has 105 GitHub stars. The repository holds 15 skills in this directory. The repository was last updated on October 5, 2026.

Source: dpc/tau on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.