.svg)
Agent-file security across repositories and workstations.
Heeler discovers and assesses instructions, skills, hooks and MCP configuration in connected repositories and on enrolled workstations. Teams can inspect risk evidence, enforce repository changes on the PR and route findings for remediation.
Agent-file discovery and assessment.
Heeler identifies risky agent files in repositories and on workstations.
Repository scans and the Workstation Sensor discover the files agents follow. Heeler assesses their content using the same scoring system across both sources.
- Discover files in repositories and on enrolled workstations
- Find instructions, skills, hooks and MCP config
- Score each file 0 to 100. Below 70 is At Risk
Repository + workstation files · At Risk first
- .claude/skills/deploy42 · Suspicious
- .cursor/mcp.json64 · At Risk
- AGENTS.md96 · Benign
Heeler evaluates agent-file changes on the pull request.
Observe, warn or block. The same check runs in CI.
Explore PR Guardrailsacme/payments-api · PR #512 · Heeler Guardrail
- Agent File Rated Suspicious · deploy/SKILL.mdBlock
- Agent File Introduced · .cursor/mcp.jsonWarn
- heelercli ci --checks agent-filesCI
Heeler routes new at-risk agent files to their owners.
A workflow posts to Slack or opens a Jira, Linear or GitHub ticket.
Explore Autonomous OperationsWorkflow · At-Risk Agent File
- WhenNew At-Risk Agent File
- SlackSafety score: 42/100
- TicketJira, Linear or GitHub
Consistent scoring across repository and workstation files.
Every score includes the findings behind it.
Static findings, model-assessed intent and external-host reputation explain each score.
.claude/skills/deploy/SKILL.md · line 14
- Static Risk · Insecure Shell PipeHigh
- LLM Risk · Remote Code ExecutionHigh
- External Risk · get.acme-tools.ioUnrecognized
Heeler assesses agent instructions and investigates observed behavior.
Use file assessments to review what an agent is instructed to do, and session evidence to investigate what it actually did.
