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 agentmods add skills/trayio/tray-plugins/build-workflownpx skills add trayio/tray-plugins --skill build-workflowgit clone --depth 1 https://github.com/trayio/tray-pluginsWhat 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 | $0.00173 | $0.06337 |
| Opus 5 | $0.00086 | $0.03168 |
| Sonnet 5 | $0.00035 | $0.01267 |
| Haiku 4.5 | $0.00017 | $0.00634 |
Grade B, and why
build-workflow scanned grade B 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
These tools come from the `tray` MCP server (configured in `~/.codex/config.toml`). Codex exposes them namespaced to that server — the exact surface form depends on your Codex version (commonly `tray.<tool>` or `tray__<t How it starts
The opening of the file, as written. The whole thing — 349 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Build Tray Workflow
Follow this process for every workflow build or modification. Do not skip the plan step.
Core principles
- Discover technically, clarify strategically. Use MCP tools to answer technical questions without asking. Surface design questions before committing to an approach.
- Understand before building. Check the existing workspace, ask 2-3 clarifying questions about intent, scope, or tradeoffs.
- Batch questions. Ask everything in one message.
- Never guess field values. When DDL can't reveal values, always ask the user explicitly.
- Salesforce custom fields (
*__c) require explicit handling. Their valid values are org-specific and unknown to the connector schema. Before setting any value on a*__cfield, either DDL its picklist, ask the user, or — if the user already supplied the value — still call out the assumption in the build summary ("Built withPriority__c = 'High'— assumes that's a valid picklist value in your org"). Building silently is a fail. Seetray-gotchasfor the rule. - Build in Tray, not local files. The source of truth is always in the Tray workspace via MCP tools — never edit local JSON directly.
- Never build workflows in parallel. Dependencies (callable workflows) must be created first.
- Chain version_ids. Every mutating tool returns a new
version_idthat must be used for the next operation. - Default to delegating connector research to the Researcher subagent. Spawn an agent matching the description "Research a Tray connector — discover version, operations, auth, required fields, DDL values, and return a step-level prescriptive result". You hand it an intent (connector + operation + the user's data flow) and it hands back a typed
propertiesblock ready to drop intoupdate_workflow. The parent agent should not calllist_connector_operationsitself — operation schemas are absorbed by the Researcher, the parent just reads back a step-input ready to use. Callinglist_connector_operationsdirectly is a fallback for trivial lookups (single connector, single small operation, no DDL fields), not the default path. - Track progress through multi-step builds (use your plan/todo facility) so dependent steps run in the right order.
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.
- 2d ago First seen · 349 lines · 173 tokens per session scan B c294b61c26aa
build-workflow is a skill published in the GitHub repository trayio/tray-plugins (9 stars, last pushed 14d ago), licensed Apache-2.0. It adds 173 tokens to every session and 6,337 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.
chat-perf
Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…