Claude Agent Team Workflows 0.1.0
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This instruction-only skill is coherent for user-approved multi-agent workflows, with notes about persistent agent-team enablement, provenance metadata mismatch, and context sharing among spawned teammates.
This appears safe to use as a workflow template if you intentionally want Claude to coordinate multiple teammates. Review the workflow scope before spawning agents, limit the files or data included as inputs, avoid secrets in team prompts, and confirm the package metadata matches the publisher/version you expected.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
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.
After setup, Claude's experimental agent-team capability remains enabled until the user removes the setting.
The skill asks for a persistent local settings change to enable experimental multi-agent functionality. This is disclosed and aligned with the skill's purpose, but it persists beyond a single workflow.
Agent Teams must be enabled. Add to `~/.claude/settings.json`: ... "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1"
Enable this only if you intend to use Agent Teams, and remove the setting later if you no longer want multi-agent workflows available.
Private material included in the workflow inputs may be passed to several spawned teammates as part of normal operation.
The workflow intentionally copies task context and generated artifacts between the lead and teammates. This is central to the skill, but it means sensitive inputs may be shared across multiple agent contexts.
Each teammate MUST receive in their spawn prompt: ... Input artifact — output from previous step (Lead must relay this) ... Handoff instruction — "Message the lead with [artifact] when done"
Before running a team workflow, define the inputs narrowly and avoid including secrets, credentials, or unrelated private data in the materials sent to teammates.
Users may have less certainty about which package version or owner record they are reviewing.
The provided registry metadata lists a different owner ID and version 1.0.0, while _meta.json lists version 0.1.0 and a different owner ID. This is a provenance inconsistency, not evidence of harmful behavior.
"ownerId": "kn76dft2yxt3vdszagzadsxvkx809fc2", "slug": "claude-agent-team-workflows", "version": "0.1.0"
Verify that the registry entry and package metadata match the expected publisher/version before installing or relying on it.