Add ai detection to include the agent or model using the cli Closes #459

[SeanAlexanderHarris]

This should help with meta monitoring of agentic use of the CLI, assist with evals, potentially power optimisation for agent use.

Appends agent/model to the User-Agent header along with Prolific CLI version.

Structure: prolific-oss/cli/, then — only when an agent or model is detected — a space-separated agent/ key/value token. is whatever version.Get() returns. This approach does rely on one of a handful of popular AI model env vars being set, or one of a couple of emerging conventional env vars being set - so it is not a guaranteed catch all, just a best attempt.

Examples

# Human-driven (no agent env var detected) — base only
prolific-oss/cli/1.0.3

# Agent-driven, branded tool (fixed slug)
prolific-oss/cli/1.0.3 agent/claude-code
prolific-oss/cli/1.0.3 agent/antigravity

# Agent-driven, generic var (forwards the env var's value)
prolific-oss/cli/1.0.3 agent/cursor        # from AI_AGENT=cursor

Tested and the new header values propagate through to internal observabilty.

[prolific-snyk]

✅ Snyk checks have passed. No issues have been found so far.

Status	Scan Engine	Critical	High	Medium	Low	Total (0)
✅	Open Source Security	0	0	0	0	0 issues
✅	Licenses	0	0	0	0	0 issues
✅	Code Security	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.

[Copilot]

Pull request overview

Adds best-effort AI agent/model detection to the Prolific CLI’s HTTP User-Agent so downstream observability can attribute API traffic to agent-driven usage, while preserving the existing prolific-oss/cli/<version> base.

Changes:

• Build and send a versioned User-Agent header from client.Execute, optionally appending agent/<name> when an agent env var is detected.
• Introduce agentenv package to detect agent environment variables and validate a safe value for headers.
• Add tests covering User-Agent composition and agent env var detection behavior.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 7 comments.

File	Description
client/client.go	Constructs the new User-Agent string (version + optional agent token) for all API requests.
client/client_test.go	Adds an HTTP test asserting Execute sends the expected User-Agent.
agentenv/agentenv.go	Implements env-var based agent detection and header-safety validation.
agentenv/agentenv_test.go	Adds unit tests for detection precedence and header-value validation.

[script-this]

Per the env-vars docs, Claude Code sets CLAUDECODE=1 (no underscore) in spawned subprocesses — not CLAUDE_CODE.

Have you tested if what's present works..? (env | grep CLAUDE in a live session)? Be good to ensure it doesn't miss primary case.

[SeanAlexanderHarris]

Ack: Tested that Claude propagates through to internal Prolific observability via the user agent header but will check that as you suggest.