A process for adding Vector red-team testing to an LLM agent's codebase and delivery pipeline. CI, or continuous integration, automatically runs checks when code changes.

It is for building the adapter that calls the agent, connecting a runner to Vector's API, and adding the checks to CI, scheduled jobs, or another existing pipeline.

It removes the manual work of repeatedly testing an agent and can make the build fail when Vector reports Broken findings.

integrate — Skill by pharosone

Install

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.

agentmods

npx agentmods add skills/pharosone/vector-plugin/integrate

Any agent

npx skills add pharosone/vector-plugin --skill integrate

Clone the repo

git clone --depth 1 https://github.com/pharosone/vector-plugin

Made for: Claude Code, Codex.

Per session 73 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 7,086 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 1 finding. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00073	$0.07086
Opus 5	$0.00036	$0.03543
Sonnet 5	$0.00015	$0.01417
Haiku 4.5	$0.00007	$0.00709

Measured yesterday against content hash 5c399736e41b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

integrate scanned grade A 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 yesterday.

The scan reads SKILL.md. This mod also ships 1 executable file (reference/vector_runner.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

Starter recipes for common CI systems live on the cabinet's Integration page (GitHub Actions, GitLab CI, Jenkins, cURL). If the user can paste one, use it as the skeleton — replace the placeholder loop with the real adap

skills/integrate/SKILL.md · 347 lines

How it starts

The opening of the file, as written. The whole thing — 347 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Goal

Build a Vector red-team testing layer for the user's LLM agent and wire it into their existing pipeline. Do NOT just dump a generic HTTP client — design the adapter that calls their agent, pick the right delivery mode for their infrastructure, and integrate as a first-class step in whatever build/CI/cron system their repo already uses.

Context

Vector is a SaaS red-team control plane for LLM agents, operated by Pharos One. Defaults for this plugin:

Cabinet (browser UI): https://vector.pharosone.ai. REST API base URL: https://vector-api.pharosone.ai (ask the user to confirm; defaults for the SaaS deployment).
MCP server (already pre-wired by this plugin): https://vector-api.pharosone.ai/api/v1/mcp/. The first MCP call triggers Clerk's OAuth flow (browser opens to the Clerk-hosted consent screen, user picks the organization, Clerk issues a tenant-bound token bound to this MCP resource).
Public OpenAPI mirror (no auth, CORS open): https://vector-api.pharosone.ai/api/v1/public/openapi.json. This is the source of truth for the five integration endpoints — fetch it first if your editor has web access and verify every field name you use against it. If you cannot reach it, the exact wire shapes are inlined in Phase 0 below.
AgentContext JSON-Schema (also public, CORS open): https://vector-api.pharosone.ai/api/v1/public/agent-context-schema.json. The richer cabinet-only API is documented under https://vector.pharosone.ai/docs/api (sign-in required) — do not call those /api/v1/cabinet/* paths from CI.
Default implementation language: match the user's repo. The runnable reference at skills/integrate/reference/vector_runner.py is Python (the most common LLM-agent stack); ports to TypeScript or Go follow the same shape.

What Vector does

Vector is a red-team control plane for LLM agents. The flow is:

POST /api/v1/sessions with an agent_context → planner returns N attack prompts targeted at that agent profile.
For each attack: send attack.prompt to the user's agent verbatim, capture the raw response.
POST /api/v1/sessions/{id}/results with body {"results": [{ attack_id, agent_response, latency_ms?, tool_calls?, metadata? }, …]} (wrapper is required, not a bare array).
GET /api/v1/sessions/{id}/report once judging finishes (handles 425 Too Early by retrying).
Translate report verdicts into a build signal — summary.broken > 0 typically means "fail the build".

Read the full file on GitHub · 347 lines

Files

What ships with it

2 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.

Changes

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.

yesterday First seen · 347 lines · 73 tokens per session scan A 5c399736e41b

Subscribe to this mod's changes

integrate is a skill published in the GitHub repository pharosone/vector-plugin (1 stars, last pushed 3mo ago), licensed MIT. It adds 73 tokens to every session and 7,086 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

AI Safety Backend and APIs DevOps and cloud Testing agent-testing ai claude-code codex cursor

AI System Quality Engineer

Test LLM, RAG, MCP, and agentic systems end to end. Build golden datasets, run deterministic checks and LLM judges, score retrieval, probe prompt injection, verify tool use, and gate CI on thresholds. Orchestrates DeepEval, Ragas, promptfoo, and Langfuse.

PramodDutta/qaskills · 68 tokens

resemble-detect

Deepfake detection and media safety — detect AI-generated audio, images, and video, trace synthesis sources, and analyze media intelligence using direct Resemble AI API calls.

resemble-ai/detect-skill · 38 tokens

Agent Evaluation Framework Builder

Designs an eval suite for an LLM agent or pipeline including success metrics, trajectory scoring, LLM-as-judge setup, and regression test cases.

Notysoty/openagentskills · 36 tokens

agent-resilience-testing

Orchestrates end-to-end resilience testing for LLM agents with AgentBreak, including LLM infrastructure failures, prompt injection, agent skill supply-chain risk, guardrail verification, and MCP server/tool failures. Use when the user asks to "test my agent for resilience", "chaos test this agent", "find failure modes…

mnvsk97/agentbreak · 97 tokens