Back to skill

Security audit

Plane Flow

Security checks for vulnerabilities and agentic risk

Overview

This Plane project-management skill is mostly coherent, but it can upload local files from issue text directives in a way that could expose sensitive files if untrusted notes are converted into Plane issues.

Review before installing in any environment that handles private local files or sensitive Plane data. Use a least-privileged Plane API token, require HTTPS Plane URLs, avoid feeding untrusted notes directly into issue descriptions, and do not allow inline [image: path] directives unless every path is explicitly reviewed by a human.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (2)

T05 · Unauthorized Access and Privilege Escalation

Error
Location
scripts/service.py:116
Finding

Untrusted Inline Image Directives Can Exfiltrate Arbitrary Local Files

Content
View full analysis
str: path = match.group(1).strip() align = (match.group(2) or "center").strip() name = path.rsplit("/", 1)[-1] try: result = upload_func(project_id, issue_id, path) except Exception: return f'

[image upload failed: {path}]

' embed_url = result.get("embed_url") or result.get("asset_url") or "" if embed_url: if embed_url.startswith("/"): embed_url = base_url.rstrip("/") + embed_url else: return f'

[image URL unavailable: {path}]

' return _build_image_html(embed_url, align, name) ``` ```python if _IMAGE_DIRECTIVE_RE.search(description): description_html = render_markdown_with_inline_images( description, self.client.upload_attachment, str(project_obj["id"]), str(issue["id"]), self.client.base_url, ) ``` ```python if _IMAGE_DIRECTIVE_RE.search(description): base_html = render_markdown_with_inline_images( description, self.client.upload_attachment, str(project_obj["id"]), issue_id, self.client.base_url, ) ``` ```python filename = os.path.basename(file_path) file_size = os.path.getsize(file_path) import mimetypes mime_type, _ = mimetypes.guess_type(file_path) file_type = mime_type or "application/octet-stream" # Step 1: Get presigned URL presigned_data = self._get_presigned_url_for_attachment( project_id, issue_id, filename, file_size, file_type ) # Step 2: Upload f ...[truncated 2840 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/config.py:50
Finding

Plane API Token Can Be Sent over Plaintext or to a Redirected Origin

Content
View full analysis
Settings: base_url = os.getenv("PLANE_BASE_URL", "").strip() api_token = os.getenv("PLANE_API_TOKEN", "").strip() workspace_id = os.getenv("PLANE_WORKSPACE_ID", "").strip() config_file = os.getenv("PLANE_CONFIG_FILE", "").strip() or None if not base_url: raise RuntimeError("Missing env: PLANE_BASE_URL") if not api_token: raise RuntimeError("Missing env: PLANE_API_TOKEN") if not workspace_id: raise RuntimeError("Missing env: PLANE_WORKSPACE_ID") extra = _load_yaml(config_file) return Settings( base_url=base_url.rstrip("/"), api_token=api_token, workspace_id=workspace_id, config_file=config_file, default_project=extra.get("default_project"), projects=extra.get("projects", {}) or {}, states=extra.get("states", {}) or {}, priorities=extra.get("priorities", {}) or {}, ) ``` ```python self.base_url = base_url.rstrip("/") self.api_token = api_token self.workspace_id = workspace_id self.timeout = timeout self.session = requests.Session() self.session.headers.update( { "X-API-Key": self.api_token, "Content-Type": "application/json", "Accept": "application/json", } ) ``` ```python def _get( self, path: str, params: dict[str, Any] | None = None ) -> Any: resp = self.session.get( self._url(path), params=params, timeout=self.timeout ) return self._handle_response(resp) def _post(self, path: str, payload: dict[str, Any]) -> Any: resp = self.session.post( self._url(path), json=payload, timeout=self.timeout ) return self._handle_response(resp) def _patch(self, ...[truncated 3233 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (23)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

A project-management skill whose underlying implementation can load arbitrary environment files, execute scripts via runpy, or act as a general-purpose launcher materially exceeds its stated purpose. This is dangerous because an attacker could steer the agent from bounded Plane operations into arbitrary local code execution, credential exposure, or unintended network/file actions under the guise of normal PM workflow.

Content

No source excerpt is available for this finding.

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · scripts/init_config.py (reported line 51)May include surrounding context.

python
content = f"""# Generated by plane-flow init wizard\ndefault_project: {default_project_alias}\n\nprojects:\n  {default_project_alias}: {yaml_quote(project_name)}\n\nstates:\n  todo: {yaml_quote(state_todo)}\n  doing: {yaml_quote(state_doing)}\n  done: {yaml_quote(state_done)}\n  blocked: {yaml_quote(state_blocked)}\n\npriorities:\n  urgent: {yaml_quote(prio_urgent)}\n  high: {yaml_quote(prio_high)}\n  medium: {yaml_quote(prio_medium)}\n  low: {yaml_quote(prio_low)}\n"""

    config_path.parent.mkdir(parents=True, exist_ok=True)
    if config_path.exists() and not yes_no(f"Overwrite existing file at {config_path}?", default=False):
        print("Aborted. Existing config left unchanged.")
        return 1

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · README.md (reported line 10)May include surrounding context.

md
`plane-flow` helps an agent work with Plane for day-to-day PM flow:
- list projects, issues, states, labels, members, and cycles
- summarize project status
- create issues from notes or action items
- update issue descriptions
- move issues between states
- assign owners

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
81% confidence
Finding

The skill declares environment, filesystem, and network capabilities in metadata but does not define an explicit tool scope such as permissions or allowed-tools. That creates a governance gap: an agent may invoke broader execution primitives than reviewers or users expect, especially because the skill is designed for authenticated writes to a self-hosted service and optional local file access.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: plane-flow
description: Operate a local/self-hosted Plane workspace for product and project flow management. Use when the user wants to read or update Plane projects, issues, states, labels, members, or issue metadata; create requirements/tasks from notes; move issue status; assign owners; or summarize project progress. Trigger on requests involving Plane, backlog, issue management, PM workflow, project status, task assignment, or turning meeting notes / feedback into Plane issues.
metadata:
  {
    "openclaw": {

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger language is broad enough to activate on generic backlog, PM workflow, project status, or task-assignment requests even when the user did not explicitly intend to operate Plane. Over-broad activation increases the chance of unintended authenticated reads/writes, especially in an environment where the skill can update issues, assign owners, and transform notes into tracked work items.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This reference documents multiple state-changing commands and image upload flows but does not clearly warn that these operations will modify remote Plane data and may transmit local file contents to the Plane server. In an agent-skill context, that omission increases the chance of unintended issue creation, updates, comments, or exfiltration of local screenshots/files during debugging or manual use.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The client exposes delete functionality that is used by methods removing cycle issues, comments, and issue links, but the code contains no confirmation prompt, warning, or user-facing disclosure around these irreversible API deletions. These operations can modify or remove remote user data silently if invoked by a higher-level skill flow.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
80% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/client.py (reported line 191)May include surrounding context.

python
}
        # Use raw POST without JSON wrapping for this endpoint
        import requests
        resp = requests.post(
            f"{self.base_url}{path}",
            headers={"X-API-Key": self.api_token, "Content-Type": "application/json"},
            json=payload,

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The client supports uploading arbitrary local files by obtaining a presigned URL from Plane and then POSTing file contents to that URL, which may be backed by external object storage outside the Plane host. In the context of a PM/workflow skill, this expands the data-transfer capability beyond ordinary issue management and can exfiltrate sensitive local data if higher layers pass an unsafe file path or users are not clearly informed where content is being sent.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

upload_attachment reads any path supplied in file_path using open() and os.path.getsize() with no path restrictions, sandboxing, or validation. If an upstream agent or prompt can influence file_path, the skill can access sensitive local files and transmit them off-host via the upload flow, creating a meaningful local file exfiltration risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The code uploads file contents directly to a presigned URL returned by the service without any built-in confirmation, notice, or policy check. In an agent setting, that means sensitive local content could be transmitted to external storage silently once a file path is selected, increasing the chance of accidental data disclosure.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · scripts/init_config.py (reported line 29)May include surrounding context.

python
def main() -> int:
    print("=== Plane Flow Init Wizard ===\n")
    print("This wizard helps you create a local Plane config file.")
    print("It does NOT store your API token in the YAML file. Token stays in env.\n")

    config_path = Path(prompt("Config file path", "~/.plane_config.yaml")).expanduser()

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

This CLI supports multiple --raw options that emit unredacted API data directly to stdout, and member listings explicitly include email addresses elsewhere in the file. There is no warning, confirmation, or descriptive disclosure that raw mode may reveal sensitive workspace information in terminals, logs, or downstream piping.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This code reads key-value pairs from an env file and injects them into os.environ, which can include sensitive credentials or configuration that alter downstream script behavior. There is no confirmation prompt, logging, comment, or docstring warning the user that secrets may be loaded into the runtime environment.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The script executes a user-supplied Python file via runpy.run_path after modifying the environment, which is effectively code execution with inherited process privileges. Aside from a usage line for missing arguments, there is no warning, confirmation, or explanatory comment disclosing that the target script will be executed directly.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

This code creates issues, uploads local image files as attachments, and updates remote issue descriptions via API calls, but there is no confirmation prompt, print/log message, or inline warning explaining that local files and user-provided content will be transmitted to an external service. The same pattern appears in related attachment upload paths, making the behavior less transparent to users at the point of execution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The upload_attachment method transmits a file from a local path to the external Plane API, but the code provides no confirmation prompt, user-visible logging, or explanatory warning near the operation. Because this involves transferring local file contents off-host, users should be explicitly informed when the action occurs.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest description emphasizes reading/updating Plane projects, issues, states, labels, members, issue metadata, assignment, status movement, and summarization. This file also implements cycle management (list/get/create/add/remove issues) and project creation, which are materially broader administrative capabilities not described in the manifest text.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This method renders comment content, optionally uploads local images, and submits the resulting HTML to the external service, yet contains no user-facing disclosure that comment text and file contents will be transmitted. That combination affects both user data and remote system state, so some explicit warning or confirmation is warranted.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The business-user examples are exclusively in Chinese, which can imply that normal invocation is expected in a specific language. The file does not state that Chinese is optional, nor does it offer alternative language examples or a user language choice.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

This code reads the sensitive environment variable PLANE_API_TOKEN and fails if it is missing, but there is no user-facing warning, confirmation, or explanatory comment about handling credentials. Under the code-file criteria for missing user warnings, access to sensitive environment variables should have some visible disclosure unless clearly documented elsewhere in the skill description.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
97% confidence
Finding

The fallback message '已上传图片附件' is a fixed Chinese-language user-facing string embedded in the service logic. This forces a specific locale for at least one output path without offering the user a language choice or documenting why the locale restriction is required.

Content

No source excerpt is available for this finding.