Back to skill

Security audit

Cleaning Maintenance

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be a legitimate TIDY integration, but it gives an agent broad control over real-world property services and long-lived credentials without enough safety boundaries.

Review this skill before installing. Use it only if you trust TIDY with property addresses, access notes, booking data, maintenance details, and vendor contact information. Avoid putting hidden-key locations or other entry secrets in notes, require human confirmation before booking, canceling, deleting, spending, or assigning concierge work, and protect or rotate any TIDY token as if it were a password.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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)

T08 · Insecure Dependencies

Error
Location
SKILL.md:39
Finding

Unpinned Global Installation of Third-Party CLI Packages

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:39
Vulnerability Type: Unpinned third-party dependency installed with global scope
Risk Level: High

Vulnerable Code Snippet

bash
brew install tidyapp/tap/tidy-request    # or: npm i -g @tidydotcom/cli

Technical Analysis

The Skill instructs users to install and execute the latest available version of a third-party CLI through either a Homebrew tap or npm. Neither installation command pins a reviewed version or verifies a checksum, package signature, or source commit.

The npm alternative installs the package globally. Package installation can execute lifecycle scripts, while the installed CLI subsequently runs with the invoking user's permissions. The Homebrew alternative similarly trusts mutable package definitions and upstream artifacts.

This creates a supply-chain trust boundary that extends beyond the audited Skill files. If the package publisher account, package registry entry, Homebrew tap, dependency graph, or release infrastructure is compromised, the command could install code materially different from what was reviewed. The package would also be positioned to access the TIDY_API_TOKEN environment variable and the documented TIDY credential file when invoked.

No evidence was found that the referenced packages are currently malicious. The vulnerability is the unsafe, unpinned installation mechanism and absence of integrity controls.

Attack Path

  1. An attacker compromises the npm publisher account, Homebrew tap, release infrastructure, or an upstream package dependency.
  2. The attacker publishes a malicious release under the package name referenced by the Skill.
  3. A user follows the documented installation command without specifying a trusted version.
  4. The package manager retrieves the attacker-controlled release.
  5. Malicious package lifecycle logic or CLI code executes with the user's local privileges.
  6. The malicious code ...[truncated 830 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin the CLI to a specifically reviewed version rather than installing the latest available release.
  2. Publish and document cryptographic checksums or verifiable package signatures for approved artifacts.
  3. Link to the canonical source repository and identify the reviewed release tag or commit.
  4. Avoid global npm installation where possible. Prefer an isolated environment, a locked project dependency, or a verified standalone binary.
  5. Document whether npm lifecycle scripts are required. If they are not, recommend installation with lifecycle scripts disabled.
  6. Use lockfiles and integrity metadata for all transitive dependencies.
  7. Establish dependency monitoring and require re-review before updating the pinned version.
  8. Advise users to run the CLI with only the filesystem and environment access required for TIDY operations.

T09 · Insecure Skill Coding Practices

Error
Location
references/authentication.md:59
Finding

Bearer Tokens Never Expire and Previous Tokens Remain Valid

Content
View full analysis

Vulnerability Details

File Location: references/authentication.md:59-61
Vulnerability Type: Insecure authentication-token lifecycle
Risk Level: High

Vulnerable Code Snippet

markdown
- Tokens **do not expire**
- Logging in again returns a new token (previous tokens remain valid)
- Store your token securely — treat it like a password

Technical Analysis

TIDY bearer tokens are documented as having no expiration, and issuing a new token does not invalidate previously issued tokens. Bearer authentication grants access based solely on possession of the token; therefore, any copied token can be replayed without additional proof of identity.

This design removes two important controls: automatic reduction of the compromise window and implicit invalidation through reauthentication. If a token is exposed through the TIDY_API_TOKEN environment variable, ~/.config/tidy/credentials, process inspection, terminal output, logs, backups, or an agent-visible response, it may remain usable indefinitely.

The impact is amplified by the breadth of authenticated API operations. The same bearer credential can access property addresses and physical-access notes and can perform state-changing actions involving bookings, reservations, maintenance tasks, and service professionals.

Reading a dedicated TIDY environment variable or credential file is necessary for authenticated CLI functionality and does not, by itself, indicate credential harvesting. The principal issue is the unlimited token lifetime and lack of automatic invalidation.

Attack Path

  1. A user authenticates and receives a TIDY bearer token.
  2. The token is stored in TIDY_API_TOKEN or ~/.config/tidy/credentials, as documented.
  3. An attacker obtains the token through local malware, an exposed configuration file, logs, backups, accidental terminal disclosure, or a compromised dependency.
  4. The attacker submits the token in an `Authorization ...[truncated 1299 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace permanent bearer tokens with short-lived access tokens.
  2. Use securely stored refresh tokens with rotation and replay detection.
  3. Revoke or rotate prior tokens when users log in again, reset a password, or report credential compromise.
  4. Provide an account interface and API for listing active tokens or sessions and revoking them individually or globally.
  5. Apply least-privilege scopes so tokens used only for availability checks cannot modify bookings, addresses, or maintenance tasks.
  6. Avoid returning raw tokens in agent-visible MCP responses where possible. Use server-managed authenticated sessions or protected secret channels.
  7. Store local credentials through operating-system credential facilities such as Keychain or a secret service. If a file is unavoidable, require owner-only permissions and prevent backup or log exposure.
  8. Redact authorization headers and tokens from application, proxy, shell, and diagnostic logs.
  9. Add anomaly detection, token-use auditing, and user notifications for new devices or unusual authenticated actions.
  10. Require explicit confirmation or step-up authentication for destructive, financially consequential, or physical-access-related operations.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (18)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

The documented DELETE /addresses endpoint enables irreversible modification of user resources, and the reference provides no mention of confirmation, scoping safeguards, or ownership-validation expectations beyond a business-rule note about active jobs. In an agent-driven context, exposing destructive endpoints without explicit human-approval guidance increases the risk of accidental or prompt-induced deletion of the wrong property record.

Content

Scanner excerpt · references/rest-api-reference.md (reported line 123)May include surrounding context.

Delete Address

text
DELETE /api/v2/addresses/:address_id

Active jobs must be cancelled first.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
88% confidence
Finding

DELETE /guest-reservations can remove reservation records, which may disrupt turnovers, scheduling, and guest operations if the wrong identifier is supplied or an agent is manipulated into acting on ambiguous instructions. The danger is amplified in this skill because reservation management is a primary function and deletions can have real operational consequences.

Content

Scanner excerpt · references/rest-api-reference.md (reported line 247)May include surrounding context.

Delete Reservation

text
DELETE /api/v2/guest-reservations/:guest_reservation_id

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
88% confidence
Finding

DELETE /tasks allows removal of maintenance/issues data, which can erase operational history or suppress unresolved problems if invoked incorrectly or maliciously through an agent workflow. Because maintenance tracking is core to this skill, loss of task records can hide safety issues, impair accountability, and interrupt remediation workflows.

Content

Scanner excerpt · references/rest-api-reference.md (reported line 290)May include surrounding context.

Delete Task

text
DELETE /api/v2/tasks/:task_id

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The skill description frames a very broad operational scope across cleanings, maintenance, pricing, professionals, and issue tracking, but does not specify guardrails for when an agent should invoke it. That increases the chance of over-triggering or misrouting user intent into external actions affecting homes or offices.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill encourages free-form natural-language commands that can trigger real-world booking and maintenance actions without defining strong confirmation boundaries, required identifiers, or approval steps. In an agent setting, ambiguous phrasing can cause unintended property actions such as creating bookings, canceling services, or reporting issues against the wrong address.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This section instructs users to submit sensitive property and operational details, including address, access instructions, and parking notes, to an external service without a warning about privacy and security implications. Exposure of access details like where a key is stored materially raises physical-security risk if mishandled, logged, or accessed by unauthorized parties.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
93% confidence
Finding

The example transmits a full physical address along with highly sensitive access instructions and parking details to an external API. Even though this is expected functionality, the inclusion of secrets such as 'Key under the mat' creates a meaningful physical intrusion risk if stored insecurely, leaked in logs, or exposed through account compromise.

Content

Scanner excerpt · SKILL.md (reported line 112)May include surrounding context.

Or via REST API:

bash
curl -X POST https://public-api.tidy.com/api/v2/addresses \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{

External Transmission

Medium
Category
Data Exfiltration
Confidence
81% confidence
Finding

This API example sends maintenance issue details tied to a property to an external service, which is a data-transfer risk because it may reveal occupancy patterns, property condition, or internal operations. While less severe than access-code disclosure, such information can still be sensitive and should not be transmitted silently by an agent.

Content

Scanner excerpt · SKILL.md (reported line 200)May include surrounding context.

Or via REST API:

bash
curl -X POST https://public-api.tidy.com/api/v2/tasks \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill describes autonomous concierge resolution that can find professionals, obtain quotes, and schedule repairs, but does not warn that downstream actions may occur automatically once enabled. In practice, this can commit the user to vendor outreach or scheduling based on incomplete or mistaken issue reports, increasing operational and privacy risk.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 · references/authentication.md (reported line 8)May include surrounding context.

Sign Up (New Account)

bash
curl -X POST https://public-api.tidy.com/api/v2/auth/signup \
  -H "Content-Type: application/json" \
  -d '{
    "first_name": "Jane",

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The documentation states that CLI credentials are stored in ~/.config/tidy/credentials but does not clearly warn about the security implications of local token storage, especially because the same file also documents that tokens do not expire and old tokens remain valid. A locally stored long-lived bearer token can be stolen by other local users, malware, backups, or misconfigured permissions and then reused indefinitely.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 · references/maintenance-workflows.md (reported line 28)May include surrounding context.

bash
# 1. Create the task
curl -X POST https://public-api.tidy.com/api/v2/tasks \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The examples instruct users to submit third-party personal contact information such as full names, email addresses, and phone numbers to an external service without any warning about consent, data minimization, or privacy obligations. In a cleaning and maintenance skill, this is especially risky because users may assume it is acceptable to upload vendor or employee contact details, creating avoidable privacy and compliance exposure.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 · references/maintenance-workflows.md (reported line 225)May include surrounding context.

bash
# All open tasks
curl "https://public-api.tidy.com/api/v2/tasks?status=open" \
  -H "Authorization: Bearer YOUR_TOKEN"

# Filter by urgency and address

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This reference explicitly instructs clients to submit raw account credentials to a login tool and shows that authentication returns an API token, but it provides no warning about secure handling, storage, redaction, or least-privilege use of those secrets. In an agent/MCP context, this is risky because users or integrators may pass credentials through LLM tooling, logs, transcripts, or third-party clients that are not appropriate places for passwords or long-lived tokens.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown reference lists delete, cancel, reschedule, create, and update endpoints that can alter or remove user data, bookings, tasks, and reservations, but it does not include warnings about irreversible or user-impacting effects. Under the markdown-file criteria, descriptions should warn when behavior could affect user data or system integrity.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 · references/rest-api-reference.md (reported line 171)May include surrounding context.

Example:

bash
curl -X POST https://public-api.tidy.com/api/v2/jobs \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 · references/rest-api-reference.md (reported line 226)May include surrounding context.

Example:

bash
curl -X POST https://public-api.tidy.com/api/v2/guest-reservations \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{