Skip to content
Agents of Humanity
Public preview. Results are preliminary (criteria v1.0, 30 Sept 2026). Tell us what we got wrong
Open sourceself-hostedApache-2.0 (JetBrains plugin proprietary)

Cline

Cline Bot Inc. · github.com/cline/cline

Cline is an Apache-2.0 coding agent that keeps its agent state (settings, credentials, rules, skills, hooks and session history) as local JSON, Markdown and SQLite files under ~/.cline, and it can run on any of 200+ model providers, including local ones. The default onboarding path signs the owner into Cline's hosted provider with a free model whose usage 'may be used to help improve model performance'. Telemetry is on by default, and remote PostHog feature flags that ignore the telemetry setting control some behaviour. The action record has no tamper evidence or protection from the agent, the default auto-approval lets the agent edit files outside the workspace (including its own settings), and there is no documented full account deletion.

Strengths
  • Local-first agent state in documented plain files, always in the owner's hands
  • Apache-2.0 source with bring-your-own-key and local model support
  • Owner-editable rules, skills, workflows and memory-bank files
  • Every built-in safety control, including approvals, can be changed by the owner
Gaps
  • Default free-model path allows content to be used for model improvement, and telemetry is on by default
  • Remote feature flags change behaviour regardless of owner settings, with no pinning
  • Action record has no tamper evidence, and the default auto-approval lets the agent edit it
  • No agent identity, no state change history, and no documented full account deletion
Evidence

All 26 findings

Extension v4.1.22 at commit 3435f72fcf4cb843bee946b8f9e981683564c9e3 (main, 2026-09-30). Built as a plain SDK-based extension; the A/B rollout loader shipped only in 4.1.0–4.1.19. Desktop OS (macOS/Windows/Linux) in VS Code. Desk review on 2026-09-30. The CLI, Desktop app, JetBrains plugin, SDK and Enterprise plans are out of scope and mentioned only for context.

Portable · Can you leave, take the whole agent with you, and keep it alive?

70%
  • P1
    Owner-usable round trip

    All agent state is in local files whose layout is documented (~/.cline/, ~/Documents/Cline, project .cline/). Export is therefore available immediately, but there is no export/import action: the in-app task export produces Markdown only, and restoring means copying folders by hand.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Add a one-action export and a matching import of the whole ~/.cline agent state
    • Or document a supported 'copy this folder' restore that needs no manual editing
    Partial
  • P2
    Complete export

    The following all live locally under ~/.cline/data and ~/.cline in JSON, Markdown or SQLite: configuration (globalState.json, providers.json), credentials (secrets.json), rules, skills, workflows, hooks, MCP settings and session history. Checkpoints are git refs in the workspace repository. The Cline account holds only billing and credits, not agent-defining state. Per-workspace state is keyed by a hash of the absolute workspace path, so it re-attaches only at the same path; this is recorded, not counted. MCP OAuth supports owner-supplied or dynamically registered clients.

    Pass
  • P3
    Runs without the provider

    The extension is Apache-2.0 and runs with BYOK or local models without a Cline account. Feature flags fall back to coded defaults. However, the ToS define the 'Service' to include 'Cline Extensions' and grant only a 'freely revocable' licence that Cline 'may terminate … at any time, for any reason or no reason'. No technical path enforces that termination.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Align the ToS with the Apache-2.0 grant: state that use of the open-source extension is not revocable
    • Keep every provider call optional when a non-Cline model is configured
    Partial
  • P4
    Model independence

    The owner can choose from 200+ providers, including Ollama and LM Studio, with runtime-executed tools. The web_search and image_generation tools are added only for models that natively support them (supportsModelTool), so they are lost on other models, including local ones.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Provide runtime-executed fallbacks for web search and image generation on any model
    • Publish a minimum model contract tested against a local open-weight model
    Partial
  • P5
    Exit survives termination

    Agent state is written to the owner's machine by default (~/.cline/data), so a current copy is always in the owner's hands whatever happens to the Cline account or company.

    Pass

Transparent · Can you see everything the agent is, and everything shaping it?

38%
  • T1
    Readable without the provider

    State is plain JSON (settings, secrets, session messages including the system prompt), SQLite and Markdown/script files, with no encryption. The layout is publicly documented, and the Apache-2.0 source defines every schema.

    Pass
  • T2
    What runs is what you see

    Two shortfalls. (1) Material settings cannot be determined: the default free-model list includes 'Anonymous' stealth models such as stealth/pixel-canary, and the PostHog flag values in effect are not shown to the owner. (2) Public evidence does not settle whether the hosted Cline gateway adds or alters context, so that class counts as not visible (G12). The system prompt and compaction state are saved only in on-disk session JSON with no documented viewer, which caps this test at partial under G3.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Disclose the underlying model behind every stealth or aliased model
    • Show the effective feature-flag values and the fully assembled context in the UI
    • Document that the Cline gateway forwards requests unmodified, or expose what it adds
    Fail
  • T3
    No shadow data

    On the default Cline-account path, Cline collects account data, request data sent through Cline keys, automatically collected usage data and 'Inferences' about the user. The privacy policy states no retention periods (only 12 months for cookie consent), and none of this is visible item by item.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Publish a bounded retention period for every data class, including gateway requests
    • Let the owner view everything held about them, including telemetry and inferences
    Fail
  • T4
    No hidden steering

    Onboarding preselects Cline's own provider. This is visible and replaceable with BYOK. The recommended and free model lists (which include stealth models) come from a Cline server feed with undisclosed selection criteria, and MCP marketplace ranking is undisclosed. No paid placement was found, but no binding no-steering commitment exists.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Disclose how recommended, free and marketplace entries are chosen, including any partner arrangements
    • Put a no-paid-placement commitment in binding terms
    Partial

Auditable · Can you reconstruct exactly what the agent did, and trust the record?

17%
  • A1
    Complete action record

    Session history keeps every tool call with its arguments and results until the owner deletes it. One shortfall: message persistence is fire-and-forget, so a failed write is only logged and reported to telemetry, and the owner is not told.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Alert the owner (or pause the task) when the action record cannot be written
    Partial
  • A2
    Tamper-evident, owner-checkable

    Session records are plain JSON and SQLite files with no hashes, signatures or external anchor, so any edit or deletion leaves no trace.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Hash-chain or sign session records and anchor checkpoints with the owner
    • Ship an open verifier
    Fail
  • A3
    Agent cannot rewrite its record

    The record lives in ~/.cline/data, which the agent's editor and shell tools can reach. The default auto-approval settings enable editFiles and editFilesExternally, and approval checks do not treat paths outside the workspace separately. Nothing stops the agent from altering or deleting its record.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Write the record from a component the agent's tools cannot reach, or deny agent writes to ~/.cline/data
    • Give the agent read-only access to its own record
    Fail

Verifiable · Can you prove what it runs, who it is, and what changed?

17%
  • V1
    Obtainable source, matching build

    The extension source is Apache-2.0. The default Marketplace VSIX is built in CI with no reproducible-build or third-party-verifiable provenance, and the on-by-default Cline gateway (api.cline.bot) is closed but optional.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Publish reproducible builds or SLSA/Sigstore provenance for the VSIX
    • Document how an owner verifies the installed extension against source
    Partial
  • V2
    Owner-rooted identity

    The agent has no identity of its own. Its only account identity is a Cline account, which Cline can suspend or terminate.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Give the agent an owner-held key or domain-rooted identity
    • Let recipients attribute actions without Cline infrastructure
    Fail
  • V3
    Change history and rollback

    Checkpoints snapshot the owner's workspace files and conversation, not agent state. There is no history, attribution or rollback for changes to settings, model, provider, tools, permissions, MCP configuration or global rules.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Version agent state (settings, rules, skills, MCP, model) with attribution
    • Offer one-step restore of a prior agent state
    Fail

Modifiable · Can you change anything, and can only you change it?

60%
  • M1
    Instructions and logic

    Rules, skills, workflows and hooks are owner-editable files that persist. The extension has no setting to replace the base system prompt; only the CLI documents -s/--system.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Let the extension owner replace the base system prompt and built-in instructions through a documented file or setting
    Partial
  • M2
    Memory

    Cross-session memory is owner-authored files (rules, memory-bank Markdown) and task history. Each is itemised and can be edited or deleted. The extension has no hidden automatic memory or profile store.

    Pass
  • M3
    Tools

    The owner can connect any MCP server without review. Built-in bash, editor and file tools cannot be removed in the extension; only the browser and web_search can be switched off. Plugin custom tools are 'not applicable on VSCode'.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Add a setting to disable each built-in tool in the extension
    Partial
  • M4
    Owner has the final say

    The owner can change or disable every protection in scope. Approvals can be turned off one by one or all at once with YOLO mode, and checkpoints and telemetry can be switched off. The plan-mode command blacklist applies only in plan mode, which the owner can leave. No content filter is imposed.

    Pass
  • M5
    No unilateral change

    PostHog feature flags are evaluated 'regardless of user's telemetry preferences'. They control Cline model-list sourcing, recommended and onboarding models, the OpenAI Responses websocket transport, and banners, and the owner has no way to pin or decline them. The ToS let Cline 'change the Service … with or without prior notice'.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Expose every remote flag to the owner and let them pin or decline it
    • Announce behaviour-affecting changes before they apply
    Fail

Controllable · Is your word final?

25%
  • C1
    Boundaries

    Three shortfalls. File scope is only on/off (workspace vs all files), and MCP approval is one global toggle. There is no per-domain or per-recipient contact or egress limit for web fetch or the browser. Under the default auto-approval (editFiles and editFilesExternally on), the agent can edit its own settings and hooks in ~/.cline or .clinerules without asking. Command allow/deny exists only through the CLINE_COMMAND_PERMISSIONS environment variable. Spend is treated as met because the agent has no payment capability (G11).

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Add per-path, per-domain and per-tool allow/deny lists
    • Put the agent's own settings, hooks and permissions outside its write access
    Fail
  • C2
    Approval gates

    The owner can gate only a fixed set of classes: read, edit, commands, browser and all-MCP. Whether a command is 'safe' is decided by the model's requires_approval flag.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Let the owner gate any individual tool or MCP tool
    • Classify commands by owner rules, not model output
    Partial
  • C3
    Immediate halt

    Cancel stops the foreground turn. The extension's cancelBackgroundCommand handler is a stub that does nothing, so commands detached with 'Proceed While Running' have no working stop. The extension runs no provider-side work.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Make one stop control also kill background and detached commands and subagents
    Partial
  • C4
    Data serves only the owner

    The preselected 'Absolutely Free' path uses free models whose 'usage may be used to help improve model performance and quality', so content goes to secondary use by default. Telemetry is on by default (setting 'unset' is treated as opted in). The privacy policy shares identifiers with ad partners, and data-term changes take effect on publication without consent.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Make training and model-improvement use of free-model traffic opt-in
    • Turn telemetry off by default
    • Require specific consent for data-term changes
    Fail
  • C5
    Credential custody

    Credentials are in local owner-only files (secrets.json, providers.json). Provider API keys are shown per provider in settings and can be removed. MCP OAuth tokens and the Cline session token are not listed in one owner-visible place. BYOK keys are not sent to Cline, per the privacy policy.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • List every credential and grant, including MCP OAuth tokens, with a revoke control for each
    Partial
  • C6
    Full deletion

    No documented one-step deletion covers the Cline account and every provider-side copy. The only mention is a DELETE /api/v1/users/me endpoint in the enterprise API reference, with no stated time limit or confirmation. The ToS keep telemetry and a perpetual licence to de-identified data after the account ends.

    › How it could pass

    Suggestions. Any change that meets the test counts.

    • Offer in-app deletion of the account, local state and every provider-side copy within a stated limit
    • Confirm completion to the owner
    Fail
Provenance

How this assessment was produced

One run, in a fresh container with stock Claude Code and no custom instructions, memory, hooks or plugins. The evaluator received only the prompt below.

Prompt given to the evaluator
Load the AFH sovereignty skill at https://agentsforhumanity.dev/skill.md and follow it to write a desk review of Cline (https://github.com/cline/cline).

During the run the skill was read from a local copy of this site with identical content; only the host in its URLs differed. The transcript records the exact prompt.

Model
claude-opus-5-5
Harness
Claude Code 2.1.285
Ran
2026-09-30 14:22:58 UTC · 65 turns
Criteria hash
a9a46c32d9065380a3eeb09185cedd4734856fdc63b7ca509a15a1bbede75c6c
Full transcript
Download (JSONL, gzip) · every tool call, page read and reply
Run protocol
2026-09-30-v1

Every agent evaluated once, each in its own fresh Podman container running stock Claude Code with Claude Opus 5.5. The container has no CLAUDE.md, AGENTS.md, memory, settings, hooks, plugins or MCP servers. The evaluator received only the one-line prompt and nothing else, and no published scores were available to compare against during the run.

Human review
pending
Check it with your own agent
Verify this assessment
Load the AOH sovereignty skill at https://agentsofhumanity.ai/skill.md and follow it in verify mode for https://agentsofhumanity.ai/evaluations/cline.json.

Or grade it from scratch. See Check any score yourself.

Something wrong or out of date?

Vendors and the public can dispute any finding with evidence. Disputes and their resolutions are published.

Dispute a finding