Back to skill

Security audit

Teambition

Security checks for vulnerabilities and agentic risk

Overview

This Teambition skill is coherent, but it needs review because it can access credentials and perform live project-management changes with broad activation and limited safety guidance.

Review before installing if this will connect to a real Teambition workspace. Use least-privilege and short-lived tokens if possible, keep .teambition-token out of source control and backups, restrict its file permissions, and require explicit user confirmation before deleting, archiving, changing members, changing permissions, or creating upload credentials.

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 (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:19
Finding

Plaintext Authentication Token Storage Without Access Controls

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 19-22
Vulnerability Type: Plaintext sensitive-data storage
Risk Level: Medium

Vulnerable Code

markdown
| File | Description |
|------|-------------|
| `{baseDir}/.teambition` | Stores the MCP Server URL and userId |
| `{baseDir}/.teambition-token` | Stores the user authentication token (optional, for operations requiring login) |

Technical Analysis

The Skill prescribes storing the user's Teambition authentication token in a predictable local plaintext file, {baseDir}/.teambition-token. It does not require restrictive file permissions, verify ownership, recommend a protected credential store, or warn against committing the file to source control.

Although the repository does not contain a token or executable code that directly leaks one, this configuration practice can expose credentials to other local users or processes with access to the base directory. Backups, indexing services, diagnostic collection, or accidental repository commits may also copy the token into less protected locations.

Attack Path

  1. A user enables authenticated Teambition operations and stores a valid token in {baseDir}/.teambition-token as instructed.
  2. The file is created with permissions inherited from the environment and without an explicit 0600 restriction.
  3. An unauthorized local user, compromised process, backup service, or source-control workflow obtains the file.
  4. The attacker extracts the plaintext token.
  5. The attacker submits authenticated requests to the configured Teambition service.
  6. The attacker accesses or modifies resources allowed by the compromised user's Teambition privileges.

This attack requires an additional avenue for reading the file or obtaining a copied version; the Skill itself does not implement credential exfiltration.

Impact Assessment

Successful exploitation compromises the Teambition authority represented by the stolen token. Depending on ...[truncated 680 chars]

Remediation
View remediation

Remediation Suggestions

  1. Store authentication tokens in an operating-system credential manager, protected MCP secret facility, or dedicated secrets manager rather than in the project directory.
  2. If file-based storage is unavoidable, create the token file atomically with owner-only permissions:
    bash
    umask 077
    install -m 600 /dev/null "{baseDir}/.teambition-token"
    
  3. Validate that the token file is a regular file owned by the current user and is not a symbolic link before reading or writing it.
  4. Reject files with group or world permissions and provide a safe command to correct them:
    bash
    chmod 600 "{baseDir}/.teambition-token"
    
  5. Add .teambition-token to .gitignore and relevant backup, indexing, and diagnostic-collection exclusions.
  6. Never print, log, transmit, or include the token in error messages and command histories.
  7. Use narrowly scoped, short-lived tokens where supported, and document token rotation and revocation procedures.
  8. Clearly separate non-sensitive endpoint configuration from credentials and document the security requirements for each file.
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The activation text is broad enough to match common conversation about projects or tasks, which can cause the skill to trigger outside explicit Teambition intent. In this skill, over-triggering is more dangerous because the skill includes read/write project-management operations and access to local config/token files, increasing the chance of unintended data access or modification.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill exposes many destructive operations such as deleting tasks, archiving projects, removing members, and modifying permissions, but does not warn that these actions can change or remove real project data. In context, this is significant because the skill is explicitly connected to a live project-management system, so unclear or accidental invocation could lead to unauthorized or unintended operational damage.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

在“确认话术”示例中,文档建议可使用 ListUserProjectsV3 查看“我参与的项目”,但前面的“可用操作”工具清单并未包含该工具。这会让技能文档对实际可用能力产生误导,属于文档意图与声明能力之间的矛盾。

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

“创建任务”示例要求先调用 GetMcpTaskPrecreate 获取字段配置,但该工具未出现在前文的可用操作列表中。文档一方面声明了一组可用操作,另一方面又依赖未声明工具完成流程,构成明显的文档与能力描述不一致。

Content

No source excerpt is available for this finding.