Security audit
Security checks for vulnerabilities and agentic risk
This is a local personal finance CLI that changes the SQLite database the user points it at, with its data-mutating behavior disclosed in the skill documentation.
Install only if you are comfortable giving this CLI write access to your local finance database. Keep backups before cleanup, auto-linking, recurring discovery, category deletion, or any command run with --yes, and use dry-run/preview commands first when available.
Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
row = cur.fetchone()
if not row:
print(f"Rule {args.id} not found")
return
rule = {
"id": row[0],
Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
row = cur.fetchone()
if not row:
print(f"Rule {args.id} not found")
return
rule = {
"id": row[0],
delete_category can reassign or delete child relationships, match rules, and manual matches, then recategorize all transactions, making it a high-impact and partly irreversible data-modifying operation. The docstring documents behavior for developers, but there is no user-facing warning, confirmation, or visible disclosure in the code path before these changes are committed.
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
if yes_flag:
return True
if not sys.stdin.isatty():
print("Error: Interactive confirmation is not available. Use --yes or -y to bypass confirmation.", file=sys.stderr)
sys.exit(1)
try:
response = input(f"{prompt_message} [y/N]: ").strip().lower()
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
# delete-account
p_del_acct = subparsers.add_parser("delete-account", help="Delete an account")
p_del_acct.add_argument("id", type=int, help="Account ID")
p_del_acct.add_argument("--yes", "-y", action="store_true", help="Bypass confirmation prompt")
p_del_acct.set_defaults(func=cmd_delete_account)
# categories
Dynamic import() can load arbitrary modules at runtime, bypassing static analysis and potentially importing malicious code.
p_stats_cat = subparsers.add_parser("stats-category", help="Total for a category (inc. children)")
p_stats_cat.add_argument("name", help="Category name")
p_stats_cat.add_argument("--month", help="Filter to YYYY-MM")
p_stats_cat.add_argument("--from", dest="from_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="Start date (YYYY-MM-DD)")
p_stats_cat.add_argument("--to", dest="to_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="End date (YYYY-MM-DD)")
p_stats_cat.add_argument("--period-type", choices=["calendar", "salary", "default"], default="default",
help="Period type: calendar, salary, or default (dynamically determined by active salary config)")
Dynamic import() can load arbitrary modules at runtime, bypassing static analysis and potentially importing malicious code.
p_stats_cat.add_argument("name", help="Category name")
p_stats_cat.add_argument("--month", help="Filter to YYYY-MM")
p_stats_cat.add_argument("--from", dest="from_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="Start date (YYYY-MM-DD)")
p_stats_cat.add_argument("--to", dest="to_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="End date (YYYY-MM-DD)")
p_stats_cat.add_argument("--period-type", choices=["calendar", "salary", "default"], default="default",
help="Period type: calendar, salary, or default (dynamically determined by active salary config)")
g_stats_cat = p_stats_cat.add_mutually_exclusive_group()
Dynamic import() can load arbitrary modules at runtime, bypassing static analysis and potentially importing malicious code.
# stats trend
p_stats_trend = subparsers.add_parser("stats-trend", help="Monthly breakdown for a category")
p_stats_trend.add_argument("name", help="Category name")
p_stats_trend.add_argument("--from", dest="from_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="Start date (YYYY-MM-DD)")
p_stats_trend.add_argument("--to", dest="to_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="End date (YYYY-MM-DD)")
p_stats_trend.add_argument("--period-type", choices=["calendar", "salary", "default"], default="default",
help="Period type: calendar, salary, or default (dynamically determined by active salary config)")
Dynamic import() can load arbitrary modules at runtime, bypassing static analysis and potentially importing malicious code.
p_stats_trend = subparsers.add_parser("stats-trend", help="Monthly breakdown for a category")
p_stats_trend.add_argument("name", help="Category name")
p_stats_trend.add_argument("--from", dest="from_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="Start date (YYYY-MM-DD)")
p_stats_trend.add_argument("--to", dest="to_date", type=lambda s: __import__('datetime').date.fromisoformat(s), help="End date (YYYY-MM-DD)")
p_stats_trend.add_argument("--period-type", choices=["calendar", "salary", "default"], default="default",
help="Period type: calendar, salary, or default (dynamically determined by active salary config)")
g_stats_trend = p_stats_trend.add_mutually_exclusive_group()
The categorize_all method clears existing rule-based category assignments for all transactions and rewrites them based on current rules, which can materially alter user financial data classification at scale. Although the docstring explains the behavior for developers, there is no user-facing warning, confirmation, or visible disclosure in code before this bulk modification occurs.
add_rule not only inserts a new rule but immediately calls categorize_all, causing bulk updates to transaction categories across the dataset. This side effect is not surfaced through any user-facing warning or confirmation in the code, so users may not realize adding one rule rewrites many records.
When a rule is deleted, remove_rule conditionally invokes categorize_all, which can reset and recompute categories for all transactions. This is a consequential, potentially disruptive operation on user data, but the code provides no user-visible disclosure or confirmation about that broad impact.
When dry_run is false, the function permanently deletes rows from id_matches and transaction_links. This is a destructive database operation, and while the docstring says it may delete rows, the code includes no confirmation prompt, print/log disclosure, or other visible warning to the user.
The method deletes transaction records when dry_run is false, which is a destructive operation affecting user financial data. Although the docstring describes optional deletion, the code provides no confirmation prompt, visible user disclosure, or explicit warning at the point of deletion.
The code allows destructive state changes to financial tracking data via remove_recurring, including hard deletion and silent soft-closing, without any built-in confirmation, authorization, audit trail, or user-warning mechanism. In a finance workflow, this can cause accidental or automated loss of historical rule state and silent changes to expected recurring-payment behavior, reducing data integrity and making mistakes harder to detect or recover from.
link_transactions can automatically create new recurring configurations, relink transactions, and assign categories based on heuristic matching, all without any visible approval gate in this module. Because these actions modify financial records and derived classifications, a bad pattern, regex, or unexpected transaction description can silently misclassify data at scale and propagate incorrect analytics.
No suspicious patterns detected.