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.
npx skills add anymouschina/TapCanvas --skill long-running-app-harnessgit clone --depth 1 https://github.com/anymouschina/TapCanvasWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/anymouschina/tapcanvas/long-running-app-harness)<a href="https://agentmods.dev/skills/anymouschina/tapcanvas/long-running-app-harness"><img src="https://agentmods.dev/badge/skills/anymouschina/tapcanvas/long-running-app-harness/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/anymouschina/tapcanvas/long-running-app-harness"><img src="https://agentmods.dev/badge/skills/anymouschina/tapcanvas/long-running-app-harness.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 206 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00038 | $0.01502 |
| Opus 5 | $0.00019 | $0.00751 |
| Sonnet 5 | $0.00008 | $0.00300 |
| Haiku 4.5 | $0.00004 | $0.00150 |
Grade A, and why
long-running-app-harness scanned grade A with 0 findings 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 12d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 221 lines — stays where its author put it; the contents beside it link to each section on GitHub.
long-running-app-harness
Use this skill when the user wants:
- a long-running autonomous app build
- planner / generator / evaluator style execution
- explicit build contracts before implementation
- skeptical QA or review loops before merging worker output
- multi-round implementation with durable artifacts
This skill is generic. It must not assume a product-specific stack, route map, or prompt pack.
Preconditions
- Load
agents-teamfirst. This skill relies onspawn_agent,wait,protocol_*, andagent_workspace_import. - Treat the persistent task graph as the durable source of truth for multi-step work.
- Use structured artifacts, not implicit chat memory, to hand off state across rounds.
- Prefer explicit failure when tools, runtime targets, or verification surfaces are missing.
Roles
orchestrator: owns the overall run, task graph, and final synthesisworker: implements one bounded slice in a private workspacereviewer: acts as the skeptical evaluator; read-only, threshold-based, evidence-first
Do not add extra roles unless the task genuinely needs different tool bounds.
Core Loop
- Create a harness run directory:
.agents/runtime/harness/<run-id>/
- Write
product_spec.json - For each round
NN:- write
round-NN-contract.json - dispatch
worker - collect staged artifacts / code handoff
- dispatch
revieweras evaluator - write
round-NN-evaluation.json - if failed: create the next round from evaluator feedback
- if passed:
agent_workspace_importand complete the task graph
- write
- Write
final-report.json
Artifact Rules
Store all harness artifacts under:
.agents/runtime/harness/<run-id>/product_spec.json.agents/runtime/harness/<run-id>/round-01-contract.json.agents/runtime/harness/<run-id>/round-01-evaluation.json.agents/runtime/harness/<run-id>/final-report.json
Do not hide important state only inside conversation history.
Product Spec Contract
product_spec.json should contain:
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.
- 12d ago First seen · 221 lines · 38 tokens per session scan A 6b9f0d559854
long-running-app-harness is a skill published in the GitHub repository anymouschina/TapCanvas (603 stars, last pushed yesterday), licensed MIT. It adds 38 tokens to every session and 1,502 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
dramake
A production workflow for turning an idea, story, script, storyboard, or existing footage into a short AI drama or film. It covers planning, generation, voice, editing, and quality checks for formats such as vertical and horizontal video.
last30Days
Resolve "last30Days" to a concrete ISO date range relative to your run time — a rolling 30-day window ending today. Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a "last 30 days" / trailing-month task…
thisWeek
Resolve "thisWeek" to a concrete ISO date range relative to your run time — this week so far (Monday → today). Returns inclusive civil dates plus exact UTC instants so you have temporal context without computing dates by hand. Read-only: no writes, no network. Use before a week-to-date task (this week's activity…
short-drama-storyboard
A workflow for turning a Chinese short-drama script and its visual facts into a shot-by-shot storyboard with frozen starting-frame prompts.
vchart-development-assistant
A coding assistant for VChart, a JavaScript chart library used to build visualizations such as bar, line, and pie charts.
md-fetch-summarize
Fetch a URL and return a concise markdown summary of its content. Read-only: no files are written; the summary is returned as output only. Use when asked to "fetch and summarize", "summarize this URL", "what does this page say", or "get the content of ". Proactively suggest when the user pastes a URL and asks what it…