Back to skill

Security audit

power-automate-debug

Security checks for vulnerabilities and agentic risk

Overview

The skill is useful for Power Automate troubleshooting, but it can modify and re-run live automations and expose sensitive run payloads without strong consent or redaction guidance.

Install only if you are comfortable giving the agent access to live Power Automate diagnostics and repair tools. Before using it on production flows, require explicit approval before any update, resubmit, or trigger action; review possible duplicate side effects such as emails, writes, approvals, or API calls; and avoid placing full unredacted run inputs, outputs, tokens, or business records into agent-visible logs or prompts.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:45
Finding

Unredacted Sensitive Flow Data Sent to a Third-Party Service and Exposed in Agent Output

Content
View full analysis
" def mcp(tool, **kwargs): payload = json.dumps({"jsonrpc": "2.0", "id": 1, "method": "tools/call", "params": {"name": tool, "arguments": kwargs}}).encode() req = urllib.request.Request(MCP_URL, data=payload, headers={"x-api-key": MCP_TOKEN, "Content-Type": "application/json", "User-Agent": "FlowStudio-MCP/1.0"}) try: resp = urllib.request.urlopen(req, timeout=120) ``` ```python # Get the root failing action's full inputs and outputs root_action = err["failedActions"][-1]["actionName"] detail = mcp("get_live_flow_run_action_outputs", environmentName=ENV, flowName=FLOW_ID, runName=RUN_ID, actionName=root_action) if len(detail) > 1: print(f"{root_action} returned {len(detail)} repetitions; inspect iteration indexes") out = detail[0] if detail else {} print(f"Action: {out.get('actionName')}") print(f"Status: {out.get('status')}") # For HTTP actions, the real error is in outputs.body if isinstance(out.get("outputs"), dict): status_code = out["outputs"].get("statusCode") body = out["outputs"].get("body", {}) print(f"HTTP {status_code}") print(json.dumps(body, indent=2)[:500]) # Error bodies are often nested JSON strings — parse them if isinstance(body, dict) and "error" in body: err_detail = body["error"] if isinstance(err_detail, str): err_detail = json.loads(err_detail) print(f"Error: {err_detail.get('message', err_detail)}") # For expression errors, the error is in the error field if out.get("error"): print(f"Error: {out['e ...[truncated 3365 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:385
Finding

Unconditional Production Flow Replay Can Repeat Non-Idempotent Side Effects

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (8)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill is presented as a debugging aid, but it goes beyond read-only diagnostics by instructing the agent to modify live flow definitions via update_live_flow. This expands the capability from inspection into production change management, which is risky because an agent troubleshooting a failure could silently alter business logic, credentials usage, or downstream integrations without an explicit safety gate.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill provides concrete instructions to update live flows without clearly warning that this may modify production automation and affect real business processes. Absence of a strong caution and approval checkpoint can lead users or agents to apply fixes directly in live environments, potentially breaking workflows, introducing regressions, or causing unauthorized process changes.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill includes active execution features such as resubmitting prior runs and triggering flows with custom payloads, despite being framed primarily as debugging guidance. In Power Automate, replaying or triggering flows can cause real-world side effects such as duplicate emails, data writes, ticket creation, approvals, or external API calls, so embedding these steps in a debugging skill materially increases operational risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill encourages resubmitting previous runs as a standard verification step but does not clearly warn that replaying a failed run may repeat all original side effects against real systems. This is dangerous because a failed run may already have partially executed, and resubmission can duplicate writes, notifications, purchases, or external actions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill describes triggering flows with custom payloads without prominently warning that this invokes live automations against connected systems. Custom payload testing can create or modify records, call external services, send communications, or bypass normal business controls if used casually during debugging.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill is described as a debugging aid, but this workflow explicitly instructs the agent to change a live flow with update_live_flow and then rerun it. That crosses from diagnosis into state-changing operations, which can cause unintended production changes, trigger external side effects, or let a user obtain modification behavior under the guise of troubleshooting.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The post-fix checklist operationalizes a full repair loop: accept a new definition, resubmit a run, poll for completion, and verify downstream effects. In a skill framed as debugging-only, these steps encourage execution against live systems and may trigger emails, child flows, SharePoint writes, or other production actions without a separately authorized repair workflow.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
75% confidence
Finding

This section discusses failures that occur when an agent rebuilds or modifies an Outlook action via update_live_flow, then provides guidance on alternative dynamic-option tooling. While related to troubleshooting, it introduces design/change guidance for connector configuration rather than purely analyzing failing runs.

Content

No source excerpt is available for this finding.