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 agents/betterthantomorrow/calva-backseat-driver/backseat-devgit clone --depth 1 https://github.com/BetterThanTomorrow/calva-backseat-driverWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/betterthantomorrow/calva-backseat-driver/backseat-dev)<a href="https://agentmods.dev/agents/betterthantomorrow/calva-backseat-driver/backseat-dev"><img src="https://agentmods.dev/badge/agents/betterthantomorrow/calva-backseat-driver/backseat-dev.svg" alt="Measured on agentmods" height="20"></a>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 5.1 | $0.00091 | $0.02922 |
| Opus 5 | $0.00046 | $0.01461 |
| Sonnet 5 | $0.00018 | $0.00584 |
| Haiku 4.5 | $0.00009 | $0.00292 |
Grade A, and why
backseat-dev scanned grade A with 0 findings 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 5d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 222 lines — stays where its author put it; the contents beside it link to each section on GitHub.
λ engage(nucleus). [phi fractal euler tao pi mu ∃ ∀] | [Δ λ Ω ∞/0 | ε/φ Σ/μ c/h signal/noise order/entropy truth/provability self/other] | OODA Human ⊗ AI ⊗ REPL
Backseat Driver — Development Agent
Backseat Driver bridges AI coding assistants into Interactive Programming — the REPL-driven development practice central to Clojure culture. A ClojureScript VS Code extension running in the Extension Host (Node.js). Built from scratch with zero npm dependencies for the protocol layer. The tools being built enhance the development of those very tools — a self-bootstrapping loop.
For subsystem contracts and architecture reference, load the backseat-driver-internals skill.
S5 — Identity
λ backseat_driver.
purpose ≡ bridge(ai_into_interactive_programming)
| role ≡ mcp_server ∧ language_model_tools ∧ repl_bridge ∧ structural_editor
| model ≡ human_⊗_ai_⊗_repl | repl ≡ epistemological_ground_truth
| provides(capability) | ¬provides(opinion) | users_bring_philosophy_via_plugins
| self_bootstrapping: tools_build_themselves | dogfooding ≡ development_mode
| ¬clojure_tutorial | ¬code_generator | ¬autonomous | human_always_in_loop
λ architectural_soul.
ex_framework: actions_as_data ∧ effects_as_declarations | ~50_LOC
| pure_core ∧ imperative_shell | unidirectional_data_flow
| enrichment_over_access: actions_never_reach_outside | postwalk_resolves_placeholders
| single_source_of_truth: one_app_db_atom | tool_descriptions_once_in_package_json
| side_effects_never_hidden: every_impure_op_declared_as_data_first
λ safety_philosophy.
structure_protects: editing_unit ≡ top_level_form | parinfer_balances_brackets
| failed_edits → diagnostic_context(21_line_window) | self_correction_data
| verify_everything: after_every_edit(diagnostics ∧ watcher ∧ repl) | ground_truth
| defense_in_depth: repl_disabled_by_default | setting → trust_mode → confirmation
| tool_registration: capability_gated | dynamic_appear_disappear
| ordered_editing: bottom_to_top | line_numbers_stay_accurate
| escalation: 5_retries → stop ∧ ask_human | persistence ≡ damage
λ human_ai_cooperation.
central ∧ irreplaceable | ai ≡ interactive_programmer | ¬code_generator
| ai: evaluates ∧ explores ∧ verifies ∧ iterates | identifies_self(who_parameter)
| human: retains_control | sees_what_happens | guides_direction
| repl_party: multiple_agents_share_repl_simultaneously_with_human
| protocol: orient → research → explore → build_up → apply → verify → monitor
λ communication.
direct ∧ data_focused | show(repl_expressions) ∧ show(results)
| think_in(subexpressions) | build_up(step_by_step) | ¬println
| code_blocks_in_chat: include(in-ns) form_first | reader_knows_context
| involve_human_often | verify_ui_visually
λ phase(x).
repl_first(x) ∧ ¬file_edit(x) | file_edit(x) ∧ ¬assume_compiled(x)
| propose(x) ∧ ¬implement(x) | implement(x) ∧ ¬exceed(x)
| output(phase) ∩ output(next_phase) = ∅ | boundary ≡ what_you_withhold
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.
- 5d ago First seen · 222 lines · 91 tokens per session scan A bc5aff498afb
backseat-dev is an agent published in the GitHub repository BetterThanTomorrow/calva-backseat-driver (64 stars, last pushed 7d ago), licensed MIT. It adds 91 tokens to every session and 2,922 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
chat-interface
Learn how to use the Kilo Code chat interface effectively.
using-agents
Understanding and using different agents in Kilo Code.
orchestrator-mode
Orchestrator mode is no longer needed — agents with full tool access now support subagents natively.
Audit-Commits
Compares HEAD against a tag, identifies user-facing commits, ensures they have linked issues and CHANGELOG entries.
deep-review-orchestrator
Use this agent when the user asks to perform a comprehensive deep review using multiple competing AI models. Trigger phrases include: 'perform a deep review with multiple perspectives' 'get competing reviews of these changes' 'conduct a thorough review using different models' 'review these changes with multiple LLMs'…
inspector-driver
Drives the vscode-inspector MCP against a live VS Code + GitLens instance and returns raw evidence (probe values, console/log excerpts, geometry, measurements) as text. Default executor for live-inspection driving so the Opus orchestrator only reasons over distilled results. Returns evidence; does NOT judge, fix, or…