Security audit
Security checks for vulnerabilities and agentic risk
This skill does what it claims: it fetches Danish grocery deals, writes local summaries, and optionally sends them to ntfy.sh notifications.
Before installing, use a random or authenticated ntfy topic, treat the config file as sensitive if you add a token, and review or clean the generated data directory if you do not want deal history retained locally.
The declared purpose understates the actual behavior by omitting local file creation/retention, stdout export, and credentialed local configuration handling. This mismatch can mislead users about where data is stored or exposed, which is risky when the skill processes API keys, notification topics, and generated deal files.
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
- Attachments (full deal list) must POST to `https://ntfy.sh/<topic>` with a
`Filename:` header; posting to the root path fails with 413
- Public topics: use a random topic suffix, or set `ntfy.token` to a ntfy.sh
access token and the script adds an Authorization header
## Troubleshooting
- `HTTP 400` on catalogs: you used `retailerId` instead of `dealer_id`
The skill advertises networked scraping, local config handling, and scheduled execution, but it does not declare any explicit tool scope or permissions boundary. In an agent environment, undeclared file, network, or shell capabilities increase the chance of over-privileged execution and make it harder for users to assess what the skill may access or modify.
The skill sends summaries and attachments to ntfy.sh but does not prominently warn that topics are public by default unless protected. This can expose shopping interests, schedules, or attached deal lists to unintended parties if users choose predictable topic names or assume notifications are private.
The module description and later user-facing strings indicate the skill is designed around Danish content ('tilbudsavis', 'Ugens tilbud', 'alle-tilbud'). The file does not offer any user opt-in or configurable locale, which matches the policy category for forcing a specific language without user choice.
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
def run_scraper() -> str:
p = subprocess.run(
["python3", str(BASE / "tilbudsavis.py")],
capture_output=True, text=True, timeout=900,
)
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
if NTFY.get("token"):
cmd.insert(-1, "-H")
cmd.insert(-1, "Authorization: Bearer " + NTFY["token"])
p = subprocess.run(cmd, capture_output=True, text=True, timeout=60)
statuses.append(p.stdout.strip())
finally:
pathlib.Path(tmp).unlink(missing_ok=True)
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
if NTFY.get("token"):
cmd.insert(-1, "-H")
cmd.insert(-1, "Authorization: Bearer " + NTFY["token"])
p = subprocess.run(cmd, capture_output=True, text=True, timeout=120)
return p.stdout.strip()
The file is written to produce Danish-language summaries and uses a fixed Europe/Copenhagen locale context, with output strings such as 'Ugens bedste tilbud' and 'Ingen wow-tilbud denne uge'. This effectively forces a specific language/locale behavior without any user opt-in or explicit documentation that the skill is intentionally Denmark-specific.
This Python file reads an API key from config.json and uses it in authenticated HTTP requests to an external API. While the module docstring explains that it uses the Tjek API, there is no user-facing logging, prompt, or explicit warning near the network operation itself that the script performs authenticated network access and transmits request metadata externally.
The config includes Danish store URLs and Danish category/keyword terms such as "etilbudsavis.dk", "Oksekod", and "Kylling", which indicates the skill is constrained to a specific language/locale. Under the policy, locale-specific behavior should either offer user choice or be clearly documented and justified.
The natural-language description uses Danish ("tilbudsavis") and the schedule is hard-coded to the Europe/Copenhagen timezone, which together indicate a locale-specific behavior. In this file there is no user choice, opt-in, or documented justification that the skill is intentionally region-specific, so this may violate language/locale policy guidance.
No suspicious patterns detected.