Feature-complete companion for the actual CLI, an ADR-powered CLAUDE.md/AGENTS.md generator. Runs and troubleshoots actual adr-bot, status, auth, config, run...
Security Analysis
high confidenceThe skill's requirements, instructions, and included diagnostic script are coherent with its stated purpose as a companion for the Actual CLI; nothing requests unrelated credentials or installs arbitrary code from remote URLs.
The skill is a companion for the 'actual' CLI and all declared requirements and behaviors align with that purpose: it requires the 'actual' binary, documents runner-specific API keys (Anthropic, OpenAI, Cursor) and includes troubleshooting and config guidance for those runners. The only small mismatch is registry metadata that lists 'Required env vars: none' while a primary credential (ANTHROPIC_API_KEY) is declared — this is minor and consistent with multi-runner behavior (the CLI can operate without an env var if a runner/binary is present).
Runtime instructions and the included diagnose.sh are limited to discovery and diagnostic actions (checking binaries, running 'actual auth', 'actual status', printing config path, grepping output files, etc.). These actions are directly relevant to installing, configuring, and troubleshooting the Actual CLI. The script is explicitly read-only and claims not to print secrets (it reports whether env vars are set and attempts to redact api_key lines). Note: 'actual config show' and similar commands may reveal config contents depending on the underlying CLI; the skill attempts to redact API keys but redaction depends on the format returned.
This is an instruction-only skill with no installer that downloads code from remote URLs. The only included executable artifact is a local diagnose.sh script. No remote download/extract/install steps are present in the skill bundle.
Requesting ANTHROPIC_API_KEY as the primary credential fits the anthropic-api runner described. References to OPENAI_API_KEY and CURSOR_API_KEY are justified because the CLI supports multiple runners. Minor inconsistency: registry summary lists 'Required env vars: none' while primaryEnv is ANTHROPIC_API_KEY — clarify which is enforced. Also note the diagnostic flow runs 'actual config show' and echoes a partial key prefix in examples (e.g., head -c 7); these are useful for troubleshooting but can expose partial secrets and rely on redaction heuristics that may not cover every key format.
The skill does not request 'always: true', has no install script that persists files or modifies other skills, and is user-invocable. It does not ask to modify other skills' configs or system-wide settings. The included diagnose.sh is read-only by design.
Guidance
This skill appears to be what it says: a helper for the Actual CLI. Before installing or running it, consider: 1) Verify you trust the Actual CLI source (https://cli.actual.ai) before installing the 'actual' binary; the skill assumes that binary is authoritative. 2) Decide where you will store API keys; the CLI can store keys in ~/.actualai/actual/config.yaml (the skill states it creates files with 0600), or you can use environment variables. 3) The included diagnose.sh runs commands like 'actual config show' and prints whether keys are set; it attempts to redact keys but redaction depends on the CLI output format — avoid running it in contexts where revealing keys (even partially) is unacceptable. 4) Clarify the registry metadata mismatch (it lists no required env vars but declares ANTHROPIC_API_KEY as primary) if you need strict policy enforcement. If you are comfortable with these trade-offs and you only want a CLI helper that runs local diagnostics and invokes the 'actual' CLI, this skill is coherent with that purpose.
Latest Release
v0.1.0
- Initial release of the feature-complete companion for the actual CLI. - Supports running and troubleshooting the actual adr-bot, covering status, authentication, config, runners, and models. - Full guidance for all five runners (claude-cli, anthropic-api, openai-api, codex-cli, cursor-cli) and model-to-runner inference. - Includes detailed operational workflows (pre-flight checks, dry-run, sync, diagnostics) and troubleshooting for all error types. - Quick reference sections for sync, runner selection, and error diagnosis. - Step-by-step install instructions for the actual CLI and fallback approaches.
More by @poiley
Published by @poiley on ClawHub
