Ask Across EverythingCAUTION
A framework for building agentic apps
Overview
A framework for building agentic apps
5977febf67f1OBSERVED · 2026-10-14What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: ask-across-everything description: >- Use when a Brain question asks across company memory plus live/app-owned data such as metrics, email threads, workspace grants, or secrets. --- # Ask Across Everything ## Rule Search Brain first, inspect `federatedCoverage`, then delegate live app-owned questions to the owning specialist agent with `call-agent`. ## Why Brain owns cited company memory: reviewed knowledge, accessible raw captures, and source records. It does not directly search Analytics, Mail/Gmail, Dispatch, or sibling app databases. `federatedCoverage` is deterministic routing metadata, not retrieved evidence. ## Workflow 1. Call `search-everything` for broad questions, or use `ask-brain` when the user asks a direct cited-memory question and wants an answer immediately. 2. Read the Brain results and `federatedCoverage.scopeNote`. 3. Inspect `federatedCoverage.delegationHints` and `federatedCoverage.discoveredAgents.agents`. `delegationHints[].target` is typed to exactly three values today: `"analytics"`, `"mail"`, `"dispatch"` (`FederatedDelegationTarget` in `server/lib/search.ts`) — there is no delegation hint for other templates yet, even if one is installed in the workspace. Each hint also carries `matchedSignals` (which words in the question triggered it) and a human-readable `reason`. 4. If the user asked for live/app-owned data and a hint matches the request, call the specialist app agent with `call-agent`. Keep the prompt narrow and include the user question plus any Brain context needed for names, dates, or citations. 5. Synthesize the final answer with clear source boundaries: - Brain knowledge/captures are cited company memory. - Specialist agent results are app-owned live data or app-native search. - Coverage metadata only explains where you routed next. 6. If no specialist agent is available, say Brain found a delegation hint but you could not query that app in this workspace. ## Routing Examples - Product decision, policy, process, or "why did we decide..." -> Brain only. Search Brain, open `get-knowledge` / `get-capture` as needed, cite evidence. - Dashboard metric, KPI, funnel, cohort, revenue, or chart -> Brain + Analytics. Use Brain for context, then `call-agent` Analytics for live dashboard/data-source facts. - Email, customer thread, sender/recipient, Gmail, inbox, or mailbox state -> Brain + Mail/Gmail. Use Brain for durable context, then `call-agent` Mail/Gmail for mailbox-native search. - Integration grants, reusable connections, secrets, approvals, recurring jobs, workspace resources, or cross-app routing -> Brain + Dispatch. Use Brain for documented decisions, then `call-agent` Dispatch for current control-plane state. ## Don't - Do not claim Brain searched sibling app databases directly. - Do not treat `federatedCoverage.delegationHints` as answer evidence. - Do not invent live metrics, mailbox state, grants, or secrets from Brain memory. - Do not put A2A calls inside Brain actions; cross-app delegation happens in the agent loop. ## Related Skills - `brain` - `delegate-to-agent` - `a2a-protocol` - `security`
Trust audit
CAUTIONgrade B · trust 89/100 Install with care. The audit found things worth knowing before you trust its output.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | NA |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (5)
.agents/skills/an
.agents/skills/content-product-development
.agents/skills/design-exploration
.agents/skills/turn-into-app
.agents/skills/visual-edit
Gates applied: no_behavioural_pass.
5977febf67f1full audit observations/trust-audit/skill/builderio__ask-across-everything.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-14 | 5977febf67f1 | CAUTION | B | 89 | first audit |
Questions
What does the Ask Across Everything skill do?
A framework for building agentic apps
Is Ask Across Everything safe to install?
With care. The audit graded it B (89/100) and found 5 things worth knowing before you trust this skill, listed below with the exact line each was found on.
What can Ask Across Everything access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
How current is this page?
The grade is for one exact copy of the source (5977febf67f1), read on 2026-10-14. The repository is watched, and a new audit runs when it changes — this is the first audit.
Add this badge to your README
Shows this page's audit result and links back to it. It updates when the audit does.
[](https://modelsatlas.com/skills/builderio/ask-across-everything/)
Not your listing, or something out of date? Claim or update your listing.