Fireflies Reference ArchitectureBLOCK
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Overview
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
4f83675ca38aOBSERVED · 2026-10-08Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned |
What 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: fireflies-reference-architecture description: >- Design a Fireflies architecture that separates identity, GraphQL reads, privileged mutations, Webhooks V2 intake, queues, sensitive data, and derived outputs. Use when performing system design or review. Trigger with "Fireflies architecture", "design Fireflies pipeline", or "Fireflies integration blueprint". allowed-tools: Read,Glob,Grep,Write,Edit argument-hint: "<repository-path> <workflow-scope>" version: 1.12.0 license: MIT author: Jeremy Longshore <[email protected]> tags: [saas, fireflies, architecture, governance] model: inherit effort: high compatibility: "Designed for Claude Code; live Fireflies work requires network access" --- # Fireflies Governed Integration Architecture ## Overview Design a Fireflies architecture that separates identity, GraphQL reads, privileged mutations, Webhooks V2 intake, queues, sensitive data, and derived outputs. ## Prerequisites - The target repository or integration path and the requested operator outcome. - The Fireflies principal, team, environment, and data classification for the work. - Current Fireflies documentation, credentials only when needed, and an accountable approver. ## Current Contract Use separate trust boundaries for bearer-authenticated outbound GraphQL, signature-authenticated inbound events, durable orchestration, meeting-content storage, and derived destinations. An event carries an identity, not permission to fetch every field. ## Authentication For authenticated operations, inject `FIREFLIES_API_KEY` from an approved secret manager and send it only as `Authorization: Bearer REDACTED_KEY` to `https://api.fireflies.ai/graphql`. Never print, commit, place in a URL, forward to a browser, or include the key in evidence. Webhook signing secrets are separate credentials and must not be reused as API keys. ## Instructions 1. Map actors, principals, environments, operations, event types, data classes, and sinks. 2. Keep API and signing secrets in separate managed stores and rotation paths. 3. Place static GraphQL documents behind a policy-aware service boundary. 4. Verify and acknowledge webhooks at the edge, then enqueue idempotent jobs. 5. Fetch only authorized fields and isolate raw, summarized, and derived data stores. 6. Add per-operation budgets, audit receipts, dead letters, deletion propagation, and rollback controls. 7. Threat-model cross-team access, replay, duplicate events, stale processing, and downstream leakage. ## Tool Discipline Use Read, Glob, and Grep to inspect code, configuration, tests, and evidence. Use Write/Edit only for approved implementation or documentation changes. Do not query Fireflies, retrieve meeting content, create an AskFred thread, upload media, change account state, replay an event, or deploy merely because this skill was invoked. ## Approval Boundaries Require approval before adding a new data sink, privileged mutation, team-wide event scope, long-term content store, or cross-tenant processing path. ## Output Return the exact operation or event surface, environment, authorization class, selected field groups, validation results, content-free metrics, decisions, and a concise pass/fail receipt. Keep secrets and meeting-derived content out of general output. ## Validation Before reporting success, rerun the smallest relevant deterministic check, compare actual state with the requested outcome and current contract, verify no secret or meeting-derived content entered logs or artifacts, and record unresolved uncertainty explicitly. ## Error Handling - Event arrives without fetch authority: acknowledge or quarantine without retrieving content. - Downstream write fails: retry idempotently within retention and budget policy. - Deletion request cannot propagate: stop new processing and escalate governance ownership. ## Examples - "Review fireflies governed integration architecture" produces a bounded plan and redacted receipt. - A request that widens access or mutates production is paused at the approval boundary. ## Resources Read [official Fireflies.ai evidence](references/official-docs.md) before relying on a field, filter, event, permission, plan limit, mutation, or processing state.
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | FAIL |
| 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 (1)
For authenticated operations, inject `FIREFLIES_API_KEY` from an approved secret manager and send it only as `Authorization: Bearer REDACTED_KEY` to `https://api.fireflies.ai/graphql`. Never print, co
Gates applied: critical_finding, no_behavioural_pass, undeclared_transfer.
4f83675ca38afull audit observations/trust-audit/skill/jeremylongshore__fireflies-reference-architecture.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | 4f83675ca38a | BLOCK | D | 69 | first audit |
Questions
What does the Fireflies Reference Architecture skill do?
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Is Fireflies Reference Architecture safe to install?
No — not without reading the findings first. The audit graded it D (69/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on.
What can Fireflies Reference Architecture access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
Which assistants does Fireflies Reference Architecture work with?
Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (4f83675ca38a), read on 2026-10-08. The repository is watched, and a new audit runs when it changes — this is the first audit.