Agency Agents 1.0.2
ReviewAudited by ClawScan on May 10, 2026.
Overview
This is an instruction-only skill with no code or credentials, but it advertises a complete 61-agent package while its own files show it is unfinished, and its autonomous workflow behavior deserves review.
Install only if you understand this appears to be an unfinished instruction/persona pack rather than a complete 61-agent suite. Verify the publisher/source, use the orchestrator with human checkpoints for important work, and avoid sharing secrets or confidential data unless you are comfortable with local saved outputs.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Users may install or pay for the skill expecting a complete agent suite, but receive an unfinished subset.
The user-facing description advertises a complete 61-agent agency, but the package's own summary says only 6 of 61 agents are complete and 55 still need conversion.
核心 Agent（6/61） ... 待转换 Agent（55/61）
Treat this as an incomplete MVP unless the publisher provides the missing agents and updates the description to match the shipped contents.
If the initial request is wrong or an intermediate agent makes a bad decision, the workflow could carry that mistake into later planning, coding, QA, or delivery steps.
The orchestrator is designed to proceed through a full multi-agent pipeline automatically after one initial command.
单个初始命令运行完整流水线 ... 无需人工干预处理错误和瓶颈 ... auto_advance: true
Use explicit approval checkpoints before file changes, publishing, deployment, business actions, or other high-impact steps.
Users cannot easily confirm the upstream source, issue tracker, or update history from the package metadata.
The package metadata points to placeholder repository URLs, which makes the skill's provenance harder to verify.
"homepage": "https://github.com/your-repo/agency-agents-openclaw#readme", "url": "git+https://github.com/your-repo/agency-agents-openclaw.git"
Verify the publisher and source before installing manually, and prefer a real repository URL with release history.
Project plans, code, business data, or other sensitive details may remain in local output folders.
The documentation says agent outputs are automatically saved locally, which can persist project content beyond the current chat.
Agent 会自动保存输出到工作区，路径：`~/clawd/agency-agents/[日期]/[项目名]/`
Review saved output folders after use and avoid providing secrets or confidential data unless you are comfortable with local persistence.