Taisly Video Publishing Kit

The documented tools focus on account discovery, video validation, and publishing. Publishing requires explicit user approval of the video, destination accounts, caption, and schedule, supporting ordinary least-privilege use and strong confirmation (least_privilege 2, user_confirmation 3). The README describes local versus remote MCP transport, the remote public-URL requirement, and local credential paths and API-key use, but gives limited detail on credential protection or data retention (data_flow_transparency 2, sensitive_data_handling 2). The supplied material contains no dependency inventory or security controls. Publishing has external social effects; confirmation and status checks help, but withdrawal or recovery is not described (dependency_security 1, external_effects 2, rollback 1). The project and package are named and registry identifiers are given, but the publisher is unverified and maintenance ownership is unclear (source_attribution 1).

The README describes broadly consistent CLI, SDK, and MCP workflows and distinguishes local files from remote URLs, but no source code or tests are supplied to verify implementation details across sources (self_consistency 2). Node 20 and npm installation are documented, but no dependency inventory, lockfile, or availability information is included (dependency_availability 1). The authentication flow says to check status and report failures, but error categories, recovery steps, and failure response formats are only lightly documented (failure_messages 1).

The material names Codex, Claude Code, Cursor, OpenClaw, and custom workflows, and covers immediate publishing, scheduling, account connection, and status checks (audience_and_scenarios 3). The exclusions and current limits define several product boundaries, though platform-specific differences and operation permission boundaries are not comprehensively described (capability_boundaries 2). The recommended path calls for validation and confirmation, and the use case is focused on social video publishing; mapping high-level requests to account selection and similar details still requires user involvement (trigger_precision 2). CLI, SDK, local and remote MCP setup are documented, including the remote URL limitation; environment coverage is mainly addressed by the Node 20 requirement and a few transport notes (environment_fit 2).

The README is organized into clear sections for setup, authentication, quick start, MCP, SDK, examples, boundaries, and limitations (information_architecture 3). Global install, npx, local development, and several MCP configurations are shown (install_notes 3). Package, command, and MCP tool names are listed, though the `integrations:*` aliases for `platforms:*` may add naming complexity (naming_stability 2). Multiple agent recipes and command examples are supplied, but there is no dedicated FAQ (examples_and_faq 3). Limits for local preflight, status endpoint behavior, remote video URLs, and planned media reuse are stated (known_limitations 3). The MIT license text is provided (license 3). Package version 0.2.9 and a syntax-check workflow are shown, but the supplied materials have no changelog or versioning policy (versioning_changelog 1). Repository and issue-tracker links exist, but no maintainer or maintenance commitment is identified, and publisher identity is unverified (maintenance_responsibility 1).

CLI and SDK examples show machine-readable JSON workflows for validation, publishing, and status checks; the MCP tool list and confirmation requirement support usable follow-on actions (output_usability 3). A shared connection to several social platforms through CLI, SDK, and MCP can reduce custom integration work, but value depends on a Taisly account, connected accounts, and its server API (marginal_value 2). Installation is described as free and setup examples are detailed; paid scaling conditions and reliance on an external service mean the practical cost-benefit is not fully established by the supplied material (cost_benefit 2).

The README, package metadata, license, and CI workflow provide traceable evidence for project identity, version, license, and syntax checking; platform coverage and service capabilities are mostly asserted by the project itself (claim_traceability 2). Documentation, package metadata, and CI corroborate names, entry points, and Node requirements, but no source code, tests, or independent material substantiates core behavior (cross_source_corroboration 2). The README distinguishes recommended workflows, stated limitations, and planned features, but marketing capability claims are not individually tied to verification evidence, so fact and inference are only partly separated (fact_inference_separation 2).