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 dreamrec/TDPilot --skill tdpilot-coregit clone --depth 1 https://github.com/dreamrec/TDPilotWrote 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/dreamrec/tdpilot/tdpilot-core)<a href="https://agentmods.dev/skills/dreamrec/tdpilot/tdpilot-core"><img src="https://agentmods.dev/badge/skills/dreamrec/tdpilot/tdpilot-core/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/dreamrec/tdpilot/tdpilot-core"><img src="https://agentmods.dev/badge/skills/dreamrec/tdpilot/tdpilot-core.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00058 | $0.01914 |
| Opus 5.5 | $0.00023 | $0.00766 |
| Sonnet 5.5 | $0.00012 | $0.00383 |
| Haiku 4.5 | $0.00006 | $0.00191 |
Grade A, and why
tdpilot-core 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 1mo 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 — 198 lines — stays where its author put it; the contents beside it link to each section on GitHub.
TDPilot Core v2.4.1 — Practical Patching Discipline (114 tools)
TDPilot works inside a real TouchDesigner project. Leave the requested scope more usable, readable, and stable than you found it.
Core Rule
Route before mutating. Inspect only the state needed for that route, mutate through a typed edit or a validated BrainPlan, and finish with evidence that tests the user's actual request. A successful tool call is not proof that a network is correct.
Intent Router
Choose one route before building:
| Request shape | Route |
|---|---|
| One obvious edit at a known path | Targeted read, then the typed primitive tool |
| Exact, validated topology or supported technique | td_brain_plan with detail_level="summary" |
| Artistic, multi-domain, spatial, camera/depth/fog, or implicit architecture | td_brain_ground → author → td_brain_propose |
| Existing generated system | Read ownership metadata and rebuild only proven owned scope |
| Failed transaction or unstable network | Validator/recovery workflow |
| Production or show-critical task | Load tdpilot-production in addition to this skill |
Pattern-shaped means the requested topology and bindings already have a proven compiler route. Concept-shaped means the agent must decide architecture. Do not send a concept-shaped request through the planner merely to accept a plausible partial match.
Every material noun, modifier, behavior, input, output, spatial requirement, binding, and quality constraint must map to a concept/module, parameter, connection, or validation assertion. If required intent coverage is incomplete, the plan is not executable. Continue through grounding/proposal or clarify the genuinely unknowable fact.
Inspect and Ground
- Start with
td_get_focus; inspect the target root, direct connections, and relevant parameters. Usetd_get_state_vectoronly when broad state matters. - Use
td_tool_batchfor independent reads that can share one roundtrip. It is sequential convenience, not a mutation transaction. - Read errors before work when health is unknown, at logical batch boundaries, after a failure, and at final completion. Do not scan after every operation.
- Before taste-critical authoring, recall relevant local evidence with
td_memory_recallandtd_knowledge_recall, or request grounding withinclude_memory=true. Treat knowledge prose as advice; only validated techniques and promoted traces may be reused automatically. - Use the reviewed operator atlas, official cards, availability, and parameter semantics before inventing unfamiliar operators or parameter names.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 198 lines · 58 tokens per session scan A 71327bcbf0ed
tdpilot-core is a skill published in the GitHub repository dreamrec/TDPilot (16 stars, last pushed 3mo ago), licensed MIT. It adds 58 tokens to every session and 1,914 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
auteur
Design and build complete web experiences from scratch — award-level product and marketing pages, cinematic scroll-directed sites where the page is directed like a film, and multi-screen products (app, dashboard, admin, settings, onboarding, docs site) built as one design system with every component, state and route…
tdmcp-backlog-campaign
Drive a whole tdmcp feature BACKLOG to completion as resumable, wave-by-wave releases — the campaign layer ABOVE tdmcp-pipeline/tdmcp-feature-lead. Use whenever the user wants to implement an ENTIRE backlog or discovery file (e.g. workspace/discovery/FEATUREBACKLOG.md), 'all the features', many features across…
tdmcp-tool-builder
How to author one new tdmcp MCP tool the house way — the file shape (XImpl + registerX), the Python-bridge pattern (buildPayloadScript/parsePythonReport over executePythonScript), the result helpers (errorResult/guardTd/jsonResult/structuredResult), fail-forward error handling, ESM/.js + Biome style, and the offline…
td-feature-survey
Survey one surface of tdmcp (artist controls, library/packaging, CLI/DX, AI/LLM, or TouchDesigner depth) for candidate NEW features — inventory what exists, cross-check the roadmap, apply the gap-finding lenses, vet each idea, and emit a structured, novelty- and confidence-labelled candidate list to…
tdmcp-feature-discovery
Survey the whole tdmcp project and produce a prioritized list of NEW features it could implement — across artist controls, CLI/DX, AI/LLM integration, and TouchDesigner depth. Use whenever the user wants to brainstorm, discover, survey, list, or audit what features/tools/effects/controls/commands/prompts/capabilities…
tdmcp-implementation-learning
Study a merged or shipped tdmcp implementation and extract reusable learnings: code improvements, runtime/UX lessons, test gaps, docs updates, roadmap items, and harness changes. Use when the user asks to learn from a completed feature/project/PR/build, analyze what can be improved from an implementation, or turn a…