Security audit

ddday

Security checks for vulnerabilities and agentic risk

Overview

This skill is a plausible work-journal and migration helper, but it can gather credentials, AI memory, environment details, and business context into portable AI-readable files without enough safeguards.

Install only after reviewing and narrowing the sensitive paths this skill can package. Before using /ddday export, /ddday snapshot, or /ddday setup, exclude credentials, .env material, and AI memory unless you deliberately need them, inspect any generated setup.sh before running it, and store any migration archive in an encrypted location.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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 (9)

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The export flow explicitly includes credential files and AI memory in a migration bundle, which materially exceeds the stated journaling/dashboard purpose and centralizes highly sensitive secrets in one portable archive. If the bundle is mishandled, copied to insecure storage, or read by another agent, it can expose authentication material, internal project context, and private history.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: The work snapshot is designed to include .env keys, credentials-related material, and system/environment details in a self-contained AI-readable file. That creates a concentrated disclosure risk because secrets and operational details are transformed into a plain consumable artifact intended for reuse across future AI conversations.

Vague Triggers

High

Confidence: 91% confidence
Finding: Mode selection is based on broad keyword matching such as 'export', 'context', or 'setup', which can trigger powerful behaviors from incidental user phrasing rather than deliberate commands. In this skill, those behaviors include packaging sensitive data and initiating restore flows, so accidental invocation materially increases risk.

Missing User Warnings

High

Confidence: 95% confidence
Finding: The migration-export section describes collecting credentials, Claude memory, and other sensitive artifacts without prominent warnings, scope limitations, or consent language. Users may reasonably assume a journaling tool exports only logs and config, not secrets and memory, leading to uninformed disclosure.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The restore workflow performs impactful system changes—installing software, copying files, rewriting paths, creating symlinks, and registering cron jobs—without clear warnings or staged consent. That can lead users to authorize broad environmental modifications they did not fully understand or intend.

Ssd 3

High

Confidence: 94% confidence
Finding: The skill's stated purpose includes creating portable bundles that let any AI agent immediately understand full work context. In practice, that framing encourages broad aggregation and transfer of sensitive user/work data across conversations and systems, expanding exposure beyond the minimum needed for journaling.

Ssd 3

Medium

Confidence: 88% confidence
Finding: The context pack is meant for new AI conversations and includes project metadata, paths, recent logs, and key file listings. Even without raw credentials, this can expose internal repository structure, sensitive filenames, operational context, and private work history to downstream systems or models.

Ssd 3

High

Confidence: 97% confidence
Finding: The export instructions call for aggregating credential files, Claude memory, cron data, skills, and environment artifacts into a single archive. Bundling these disparate sensitive materials together creates a high-value target that can enable account compromise, surveillance of workflows, and replication of the user's environment by unauthorized parties.

Ssd 3

High

Confidence: 98% confidence
Finding: The work snapshot explicitly instructs compilation of .env keys, business data, AI memory, logs, credentials, and system details into a self-contained AI-readable artifact. This is dangerous because it converts multiple classes of sensitive information into a single highly accessible document optimized for reuse and disclosure.

Static analysis

No suspicious patterns detected.