Kilo Code CLI

# Fix list: Kilo Code CLI

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/kilo-code-cli, the October 2026 research run, assessed 8 October 2026. Grade BB, 75.4 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Kilo Code CLI: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Payments & pricing, 50 out of 100, up to 6.3 more on the total

Why it scored 50: No payment protocol (0). Scored on Kilo's paid options, as with Cline. Plan prices are public (Kilo Pass from $19 a month, Teams $15 a user a month, Enterprise by quote) and the gateway bills provider rates plus a 5 per cent fee on credit purchases, but the pricing page carries no per-token list (10). The CLI is MIT-licensed and free, and the pricing page says the free tier needs no card (20). Your own provider key or a local model works with no Kilo account, so an agent can install and run it without a sign-up (20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 2. Security & auth, 66 out of 100, up to 6 more on the total

Why it scored 66: Harness reading of the framework checklist. 30 for what leaves the machine by default, 20 for approvals and sandboxing, 15 for prompt-injection posture, 15 for audit and 20 for the security programme. Telemetry is on by default and goes to PostHog in the United States, the source links the machine identifier to the account email after a Kilo sign-in, and it is turned off with `experimental.openTelemetry` or `KILO_TELEMETRY_LEVEL`. Provider keys stay on the machine (14). Shell commands ask by default outside a built-in allow list, reads of `.env` files and paths outside the project ask, and edits inside the project run without asking. An operating-system sandbox (Seatbelt on macOS, Bubblewrap on Linux, none on Windows) limits writes and blocks the network, refuses to run when it can't be enforced, and is off by default. `--auto` approves everything not explicitly denied (15). The sandboxing page covers prompt injection and its limits, project config can't weaken the sandbox, and `{env:VAR}` is resolved only in trusted config (11). Session export, `kilo stats` and OTLP export of traces and logs, with audit logs on the Enterprise plan (12). A valid security.txt (expires 2026-12-31), a disclosure policy that cites HackerOne's standards and safe harbour, CodeQL in CI and no published repository advisories, while SECURITY.md still says the CLI has no sandbox (14).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 3. Agent ergonomics, 72 out of 100, up to 4.6 more on the total

Why it scored 72: Framework reading, adapted to a harness driven by a pipeline. Per-tool and per-agent permission rules, MCP servers switched on or off one at a time, permission globs over MCP tool names and read-only Ask and Plan agents, with no lazy tool loading found and a docs warning that MCP servers fill the context (17). Exit code 124 is documented for a timeout and a `doom_loop` check stops repeated failures, but `kilo run` lists no flag for a turn, time or cost limit (11). Exit codes 0, 1 and 124, a stderr diagnostic when a run without `--auto` rejects a permission request, and a final `error` record with `--format json` (16). `--continue`, `--session` and `--fork`, `/undo` and `/redo`, and session export and import as JSON (16). `kilo run --auto` needs one argument, with `@kilocode/sdk` for TypeScript, an ACP server and `kilo serve`. We found no second-language SDK on a registry (12).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 4. Reliability, 85 out of 100, up to 3 more on the total

Why it scored 85: Local-package reading. `@kilocode/cli` on npm, a Homebrew tap, an AUR package, an install script and release binaries for Linux, macOS and Windows on x64 and arm64, with baseline and musl builds named (20). Public CI, and the 40 most recent completed workflow runs on main on 8 October 2026 all succeeded, the test, typecheck and VS Code test workflows among them (25). 448 open issues and 147 open pull requests, a workflow that closes issues after 60 quiet days and pull requests after 30, and open reports from 7 October of a turn that hangs after the loop exits (#14876) and of exhausted 503 retries ending a run with no visible error (#14877) (15). The CLI changelog is generated by Changesets with Minor and Patch headings and no dates or breaking-change headings, and the end of the fallback to `.opencode` configuration is recorded in the docs, not under a version (10). 7.8.8, past 1.0 (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 5. Transparency & trust, 76 out of 100, up to 2.1 more on the total

Made of editorial 61, provenance 90.

Why it scored 76: MIT, with opencode's copyright kept in the licence file (30). The privacy policy (29 May 2026) names Kilo Code Inc. as controller, says data is stored in the United States and lists Stripe, OpenRouter, model providers, AppsFlyer and Sentry, but it doesn't mention the CLI's PostHog telemetry. PRIVACY.md in the repository is dated March 2025 and is silent on telemetry too. The terms (30 September 2026) are a contract with Anaconda, Inc. and grant a perpetual licence to use uploaded Customer Data to improve the services (13). `kilo console` is marked deprecated with no removal date, and updated terms apply after 30 days. No written deprecation policy found (8). The docs say telemetry is on by default and name the opt-out key in two places, without listing what is collected or where it goes (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: kilo.ai, registered 2017-12-16 (8 years) (11 of 15)
- Terms of service: read, states 6 of the 7 things a reader expects, and has 2 clauses that cost points (5.1 of 10)

## 6. Schema & documentation, 90 out of 100, up to 1.6 more on the total

Why it scored 90: Framework reading. A JSON Schema for `kilo.jsonc` at `https://app.kilo.ai/config.json`, an OpenAPI file for the `kilo serve` API at `packages/sdk/openapi.json` and a CLI reference generated from the command definitions (25). `llms.txt` on kilo.ai and on the docs, the second listing a raw Markdown address for every page (10). The sandboxing page says when to use the sandbox, what it does not protect and how it differs from permissions (16). Permissions are `allow`, `ask` or `deny` with glob rules, `sandbox.network` and `--format` are enums, though the schema served on 8 October had no `sandbox` key (13). Config and CI samples and exit codes 0, 1 and 124, but we found no reference for the records `--format json` writes (11). A versioned changelog and dated GitHub releases (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 7. Maintenance & community, 87 out of 100, up to 1.1 more on the total

Why it scored 87: 7.8.8 on 7 October 2026 (30). 37 stable releases since 10 July 2026, JetBrains tags aside (20). The newest issues carry area and triage labels from automation, but the 15 newest open issues had no or one comment each on 8 October, against 448 open issues and 147 open pull requests (15). `@kilocode/sdk` and `@kilocode/plugin` are at 7.8.8 with the CLI, and the VS Code extension ships from the same release workflow (13). CI, CodeQL and Dependabot security updates run, with version updates switched off because most dependencies arrive through merges from opencode (9).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- The terms of 30 September 2026 are a contract with Anaconda, Inc. on behalf of itself and affiliates including Kilo Code Inc., and trust.kilo.ai redirects to trust.anaconda.com. We didn't read an announcement of the relationship, and the lead named only Kilo Code as vendor
- Whether the SOC 2 Type 2 report listed on Anaconda's trust centre covers Kilo's services. The page has a Kilo Code section we couldn't read in full
- Which properties the PostHog events carry. We read the event names and the client, not every call site
- Whether Kilo runs a paid bug bounty. The security page cites HackerOne's standards and names no rewards
- Unchecked: the Kilo Pass Pro and Expert prices, which the pricing page draws by script. Only the $19 entry price was in the page text
- Unchecked: where exit code 124 is set. The docs list it and `kilo run` shows no timeout flag
- Unchecked: the first release date of `@kilocode/cli` on npm, and whether a Python SDK is published (the repository has a publish workflow for one)
- The repository is a fork of opencode. How quickly upstream security fixes reach Kilo releases wasn't measured

## Weaknesses

- Telemetry to PostHog is on by default, and the privacy policy and PRIVACY.md don't mention it
- The sandbox is off by default and unavailable on Windows
- `kilo run` lists no flag for a turn, time or cost limit
- The terms grant a perpetual licence to use uploaded Customer Data to improve Kilo's services
- SECURITY.md still says the CLI has no sandbox, and the changelog has no dates or breaking-change headings

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Set `sandbox.enabled` to true in the global `kilo.jsonc` before `kilo run --auto`. `--auto` approves every permission request not explicitly denied
- Set `experimental.openTelemetry` to false, or `KILO_TELEMETRY_LEVEL` to a value other than `all`, to stop telemetry
- Add needed hosts to `sandbox.allowed_hosts` in global config. MCP tool calls are unavailable while network restriction is on
- Without `--auto` a non-interactive run rejects every permission prompt and exits 1
- Put provider keys in global config or the environment. `{env:VAR}` in a project `kilo.json` is ignored

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.