T09 · Insecure Skill Coding Practices
Warning
- Location
SKILL.md:49- Finding
Database Credentials Exposed Through Command-Line Arguments
- Content
View full analysis
- Remediation
View remediation
Security audit
Db Explorer
Security checks for vulnerabilities and agentic risk
Overview
This database skill is not malicious, but it needs Review because it includes sensitive database access, exports, restores, and migration commands with some unsafe examples and incomplete guardrails.
Install only if you are comfortable giving the agent database connection details and letting it suggest export, backup, restore, and migration commands. Use least-privileged read-only credentials by default, avoid production restores/imports unless you explicitly approve the exact command and target, do not put real passwords in command-line arguments, avoid broad Redis `KEYS *`, and write exports/backups only to private, access-controlled paths with cleanup.
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 (3)
T09 · Insecure Skill Coding Practices
Warning
- Location
SKILL.md:121- Finding
Predictable Shared Temporary Files Used for Database Exports and Backups
- Content
View full analysis
/tmp/export.csv # JSON (PostgreSQL) psql "$CONN" -t -c "SELECT json_agg(t) FROM (SELECT * FROM table_name LIMIT 100) t;" > /tmp/export.json # SQLite to CSV sqlite3 /path/to/db.db ".mode csv" ".headers on" ".output /tmp/export.csv" "SELECT * FROM table_name;" ".quit" ``` ```bash # Backup sqlite3 /path/to/db.db ".backup /tmp/backup.db" # Or just copy cp /path/to/db.db /tmp/backup_$(date +%Y%m%d).db ``` ```bash # PostgreSQL to CSV to MySQL psql "$PG_CONN" -c "\copy table_name TO '/tmp/export.csv' WITH CSV HEADER" mysql "$MYSQL_CONN" -e "LOAD DATA LOCAL INFILE '/tmp/export.csv' INTO TABLE table_name FIELDS TERMINATED BY ',' ENCLOSED BY '\"' LINES TERMINATED BY '\n' IGNORE 1 ROWS;" ``` ### Technical Analysis The commands use fixed or predictable names under the shared `/tmp` directory. Examples include `/tmp/export.csv`, `/tmp/export.json`, `/tmp/backup.db`, and a filename based only on the current date. Depending on operating-system configuration, client behavior, existing file ownership, and the invoking user's `umask`, these patterns can create the following risks: - A local attacker may anticipate the output filename and prepare a symbolic link or other filesystem object before export. - Redirection or file-opening operations may follow a symbolic link and overwrite another file writable by the victim. - Sensitive exports may be created with permissions that allow other local users to read them. - Repeated executions can overwrite previous exports or consume attacker-controlled input left at the same path. - The migration workflow trusts a predictable int ...[truncated 1845 chars]- Remediation
View remediation
T09 · Insecure Skill Coding Practices
Warning
- Location
SKILL.md:70- Finding
Unbounded Redis Key Enumeration Can Block Production Services
- Content
View full analysis
- Remediation
View remediation
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
- 95% confidence
- Finding
The activation triggers are broad phrases like 'check the database' or 'show me the data,' which can match many ordinary requests and cause the skill to activate with minimal context. In a skill that contains credential handling, schema enumeration, export, backup, restore, and migration guidance, over-triggering increases the likelihood of unnecessary access to sensitive systems or escalation from informational requests to operational actions.
- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The safety section promises read-only defaults and explicit confirmation for writes, yet later sections provide restore/import commands with no required confirmation workflow, dry-run, or rollback guidance. This contradiction is dangerous because downstream agents may follow the later concrete commands instead of the higher-level safety rules, leading to unintended data modification or loss.
- Content
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 89% confidence
- Finding
The export and backup guidance writes data directly to predictable filesystem paths such as /tmp and describes bulk extraction without prominently warning about sensitive data exposure, retention, access permissions, or overwrite risks. In real environments this can leak regulated data to shared locations or create untracked plaintext dumps that persist beyond the task.
- Content
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The skill is presented as a database exploration/query tool, but it also includes restore, migration, and import commands that can directly modify or overwrite production data. That scope mismatch increases the chance an agent will invoke destructive operations under a user request that sounded read-only, especially because the documented commands are immediately executable and not wrapped in approval gates.
- Content
Static analysis
No suspicious patterns detected.