Back to skill

Security audit

tella-cli

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a real Tella management skill, but its setup can execute mutable remote installer and npm code and alter the user's command environment.

Install only if you trust the setup chain and are comfortable with a CLI that can read, edit, share, export, and delete Tella workspace content. Prefer preinstalling Bun and api2cli from pinned, verified sources, avoid running the remote installer blindly, and treat stored API tokens and webhook signing secrets as sensitive.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:47
Finding

Remote installer is downloaded and executed without integrity verification

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 47
Vulnerability Type: Remote payload retrieval and execution
Risk Level: High

bash
bun --version || curl -fsSL https://bun.sh/install | bash

Technical Analysis

The setup instruction pipes content retrieved from https://bun.sh/install directly into bash. The downloaded payload is not pinned to an immutable version, inspected before execution, or verified against an expected cryptographic digest or signature.

Consequently, the code executed during installation can change after the Skill has been reviewed. HTTPS protects the connection in transit under normal circumstances, but it does not protect against compromise of the hosting service, DNS or certificate infrastructure, the upstream publishing account, or the installer itself.

Installing Bun is relevant to the documented setup process, but immediate execution of mutable remote content exceeds the minimum privilege necessary. A safer installation flow can download a fixed release artifact and verify it before execution.

Attack Path

  1. An attacker compromises the remote installer, its hosting infrastructure, or an authorized upstream publishing account.
  2. The attacker modifies the response served by https://bun.sh/install.
  3. A user or agent follows the Skill setup instructions on a host where Bun is unavailable.
  4. Because bun --version fails, the shell evaluates the right side of ||.
  5. curl retrieves the attacker-controlled response and streams it directly into bash.
  6. The payload executes with all permissions available to the user running the agent or setup command.

Impact Assessment

Successful exploitation permits arbitrary command execution in the invoking user's security context. The payload could read or alter user-accessible files, inspect environment variables and credentials, modify shell configuration, install additional executables, or access authenti ...[truncated 463 chars]

Remediation
View remediation

Remediation Suggestions

  • Remove the curl | bash installation pattern.
  • Direct users to Bun's official installation documentation or a trusted operating-system package manager.
  • If automated installation is required, pin an exact Bun release and download its immutable release artifact separately.
  • Verify the artifact using a hardcoded SHA-256 digest or a valid upstream cryptographic signature before installation.
  • Store the downloaded file temporarily, allow inspection, and execute it only after verification succeeds.
  • Run installation with an unprivileged account and avoid requesting elevated permissions.
  • Document the files and environment settings that installation will modify.
  • Prefer failing with a clear prerequisite error when Bun is absent rather than automatically executing remote code.

T08 · Insecure Dependencies

Error
Location
SKILL.md:48
Finding

Unpinned npm CLI dependency can execute mutable supply-chain code

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 48-49
Vulnerability Type: Insecure dependency execution
Risk Level: High

bash
npx api2cli bundle tella
npx api2cli link tella

Technical Analysis

The setup instructions invoke api2cli through npx without specifying an exact package version or integrity constraint. Depending on the local environment and npm behavior, npx may resolve and download the current package version from the configured npm registry before executing its CLI entry point.

This makes the effective executable mutable after review. A compromised upstream package, maintainer account, registry response, or locally configured registry could cause attacker-controlled package code to run. The subsequent link operation is also documented as modifying the user's command search path, increasing the importance of establishing the provenance of the installed executable.

The Tella Skill legitimately requires a supporting CLI, but executing an unpinned package is not the least-risk method of providing that dependency. No lockfile, package manifest, integrity digest, or trusted package version is included in the reviewed project.

Attack Path

  1. An attacker compromises the api2cli package, an authorized publisher account, or the dependency source configured for npm.
  2. A malicious version is published or supplied as the version selected by unqualified npx api2cli.
  3. A user or agent runs the documented setup commands.
  4. npx resolves, downloads, and executes the malicious package in the invoking user's context.
  5. Malicious package initialization or CLI code runs before or alongside the requested bundle and link behavior.
  6. The code may then alter generated CLI artifacts or path-related configuration so later tella-cli invocations execute attacker-controlled logic.

Impact Assessment

Exploitation can provide arbitrary code execution with the permissions o ...[truncated 1084 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin api2cli to a reviewed exact version rather than resolving the latest available release.
  • Manage the dependency through a committed package manifest and lockfile with integrity metadata.
  • Use an execution mode that refuses implicit installation when the expected package is not already present.
  • Verify the resolved package name, version, source registry, and integrity before executing it.
  • Review the package's lifecycle scripts and transitive dependency tree.
  • Prefer a signed, versioned standalone release artifact where available.
  • Run bundling in an isolated, unprivileged environment without Tella tokens or unrelated user credentials.
  • Do not automatically modify the user's PATH; instead, display the exact proposed change and require informed user approval.
  • Ensure generated executables are written to a controlled location and verify their ownership and integrity before subsequent use.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (7)

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

The setup instructions fetch and immediately execute a remote install script via curl -fsSL https://bun.sh/install | bash. Piping network content directly to a shell bypasses normal review and integrity checks, so a compromised host, MITM in a weak environment, or malicious script update could execute arbitrary code on the machine running the skill.

Content

Scanner excerpt · SKILL.md (reported line 47)May include surrounding context.

If tella-cli is not found, install and build it:

bash
bun --version || curl -fsSL https://bun.sh/install | bash
npx api2cli bundle tella
npx api2cli link tella

Chaining Abuse

High
Category
Tool Misuse
Confidence
98% confidence
Finding

The | bash construct is dangerous because it turns remote content into immediate shell commands with no validation step. In the context of an agent skill, this is especially risky because automation may execute setup instructions non-interactively, magnifying the blast radius of a compromised installer.

Content

Scanner excerpt · SKILL.md (reported line 47)May include surrounding context.

If tella-cli is not found, install and build it:

bash
bun --version || curl -fsSL https://bun.sh/install | bash
npx api2cli bundle tella
npx api2cli link tella

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger description includes broad phrases like 'screen recording', 'video transcript', and 'playlist of videos', which can cause the skill to activate for generic requests not specifically about Tella. Over-broad routing can lead to unintended tool use, unnecessary data exposure to this integration, or destructive actions in the wrong domain.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding

The skill instructs use of npx api2cli bundle tella without pinning an exact package version, so execution depends on whatever package version is current at runtime. That creates a supply-chain risk: a compromised, typosquatted, or breaking upstream release could execute unreviewed code during skill setup.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding

The npx api2cli link tella command also invokes an unpinned package, which means the installed CLI behavior and any code run during linking can change over time without review. In an agent skill context, dynamic bootstrap steps increase exposure to supply-chain compromise and reduce reproducibility.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · SKILL.md (reported line 74)May include surrounding context.

tella-cli auth remove # Delete stored token

text

Token is stored at `~/.config/tokens/tella-cli.txt` (chmod 600).

## Resources

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The webhooks section advertises get-secret <id> for retrieving signing secrets and discusses recent delivery/message inspection without any warning about treating the secret as highly sensitive. In an agent-operated environment, this increases the chance that secrets are fetched, logged, echoed to chat, or mishandled by downstream automation.

Content

No source excerpt is available for this finding.