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 instructions/olo-dot-io/uni-cli/agents-mdgit clone --depth 1 https://github.com/olo-dot-io/Uni-CLIWhat 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 | $0.02029 | $0.02029 |
| Opus 5 | $0.01014 | $0.01014 |
| Sonnet 5 | $0.00406 | $0.00406 |
| Haiku 4.5 | $0.00203 | $0.00203 |
Grade A, and why
Uni-CLI AGENTS.md 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 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
Before raw browser tools, OpenCLI, curl, computer-use, or an external CLI, run How it starts
The opening of the file, as written. The whole thing — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Uni-CLI
AI agents: install with
npm install -g @zenalexa/unicli. Discover withunicli search "intent". Run withunicli <site> <command>. Output is Markdown when piped, JSON via-f json. Errors are structured envelopes on stderr. Anauth_requiredenvelope means the site needs cookies — rununicli auth setup <site>once and retry. Adapter is broken? Readunicli repair <site> <command>.
Every structured response can carry meta.update. Persistent non-interactive
Agent installations schedule its exact release automatically and report state
under automatic_update. Confirm with unicli --version before retrying a
version-sensitive task. Use unicli upgrade --no-auto-update for explicit
approval. Interactive users can choose Y or N with unicli upgrade.
Open Agent-Computer Interface runtime for real software. Rank operations by
intent, explicitly select one with a declared substrate, inspect its structured
result, and repair supported drift paths. Operation-specific evidence is
optional. YAML adapters stay agent-readable and locally repairable. See
docs/BENCHMARK.md for measured Uni-CLI call costs.
Agent Routing Rule
Before raw browser tools, OpenCLI, curl, computer-use, or an external CLI, run
unicli search "<intent>" or unicli list --site <site>. Read structured
errors and run unicli repair <site> <command> before changing tools.
For auth and browser failures, read unicli browser profiles --json and
unicli browser doctor --json. Browser work starts in the background.
Foreground control requires unicli browser --focus start. CDP uses a
Uni-CLI automation profile because Chrome 136 and managed policy can block the
default profile. Empty profiles require an explicit ephemeral session. Follow
checks[*].next_step from doctor. Use unicli browser doctor --repair only
for the windowless broker.
Always-on writing rule
Read skills/human-writing/SKILL.md before every user-facing reply and prose
artifact. It is the primary Chinese and English writing policy. Its current
discussion scan and hard bans are mandatory. Load references only when the
task calls for them. Run npm run prose:check -- <path> on authored files and
apply the same scan manually to chat replies.
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 · 160 lines · 2,029 tokens per session scan A 25b5d0ae3e39
Uni-CLI AGENTS.md is an instructions file published in the GitHub repository olo-dot-io/Uni-CLI (270 stars, last pushed 12d ago), licensed Apache-2.0. It adds 2,029 tokens to every session, about $0.0101 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 2026-08-30.
Other instructions, from other repositories
empir3-bridge AGENTS.md
Instructions for empir3hq/empir3-bridge, covering empir3 bridge — agent & contributor guide, what lives here, before you edit, run locally and release / publish.
empir3-bridge CLAUDE.md
Instructions for empir3hq/empir3-bridge, covering empir3 bridge — agent & contributor guide, what lives here, before you edit, run locally and release / publish.
empir3-bridge GEMINI.md
Instructions for empir3hq/empir3-bridge, covering empir3 bridge — agent & contributor guide, what lives here, before you edit, run locally and release / publish.
browser-use CLAUDE.md
Claude Code instructions for browser-use/browser-use, covering claude.md, high-level architecture, core components, event-driven browser management and cdp integration.
obscura AGENTS.md
Instructions for h4ckf0r0day/obscura, covering agents.md, build, rendering and stealth, no rendering, with rustls or stealth and test.
browser-harness AGENTS.md
Instructions for browser-use/browser-harness, covering code priorities and contributing.