task-viewer

A read-only web dashboard that displays Claude Code tasks as a live Kanban board, including pending, active, and completed work.

In plain words
What is it for?
Use it to launch the task viewer in a browser and observe tasks, dependencies, timelines, and activity. It does not create or edit tasks.
Why use it?
It gives developers a visual way to inspect task progress across sessions without manually opening task files.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/surya8991/agentmaster/task-viewer
Any agent
npx skills add Surya8991/AgentMaster --skill task-viewer
Clone the repo
git clone --depth 1 https://github.com/Surya8991/AgentMaster

Made for: Claude Code, Codex.

Per session 80 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 630 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00080 $0.00630
Opus 5 $0.00040 $0.00315
Sonnet 5 $0.00016 $0.00126
Haiku 4.5 $0.00008 $0.00063

Measured 2d ago against content hash 5c249dcf398d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

task-viewer scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

`status`: `curl -s -o /dev/null -w '%{http_code}' http://localhost:3456` (try the reported port) — 200 means running.
.antigravity/skills/task-viewer/SKILL.md · 45 lines

What it actually says

Launches and manages the claude-task-viewer dashboard — a real-time Kanban board that reads ~/.claude task JSON and serves it in the browser. Observation only: Claude Code controls task state, this just visualizes it.

Resolve the app directory

The app lives outside ~/.claude/skills (it's a full Node app, not a skill). Its location is stored in a config file so this skill stays portable:

CFG="$HOME/.claude/.agentmaster-cache/task-viewer.path"
APP_DIR="$(cat "$CFG" 2>/dev/null)"

If $CFG is missing or $APP_DIR doesn't contain server.js, tell the user the app isn't installed/registered and ask for its path, then write it back: printf '%s' "<path>" > "$CFG". Do not guess absolute paths.

Commands

Default (no arg) or start:

  1. Read $APP_DIR as above.
  2. Ensure deps exist: if $APP_DIR/node_modules is missing, run npm install in $APP_DIR first.
  3. Start the server in the background and open the browser: PORT=<port-if-given> node "$APP_DIR/server.js" --open
    • Default port is 3456; it auto-increments if busy (unless PORT is set).
    • Run it as a background process so it keeps serving after this turn.
  4. Report the URL it printed (http://localhost:<port>).

status: curl -s -o /dev/null -w '%{http_code}' http://localhost:3456 (try the reported port) — 200 means running.

stop: kill the running server process.

  • Windows: taskkill //F //IM node.exe is too broad — prefer finding the PID bound to the port. If unsure, tell the user which PID/port and let them confirm.
  • Unix: pkill -f "server.js" scoped to this app.

Notes

  • The viewer reads task data from $CLAUDE_DIR (defaults to ~/.claude). Override with CLAUDE_DIR=<path> or --dir <path> if the user keeps tasks elsewhere.
  • It's read-only over tasks; never claim it can create/modify/delete task state on Claude's behalf.
  • This is a long-running server. Don't block the conversation waiting on it — start it detached and move on.
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 45 lines · 80 tokens per session scan A 5c249dcf398d

Subscribe to this mod's changes

task-viewer is a skill published in the GitHub repository Surya8991/AgentMaster (2 stars, last pushed 1mo ago), licensed MIT. It adds 80 tokens to every session and 630 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.