Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add scholarly360/owasp-top10-web-skills --skill software-data-integrity-failuresgit clone --depth 1 https://github.com/scholarly360/owasp-top10-web-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/scholarly360/owasp-top10-web-skills/software-data-integrity-failures)<a href="https://agentmods.dev/skills/scholarly360/owasp-top10-web-skills/software-data-integrity-failures"><img src="https://agentmods.dev/badge/skills/scholarly360/owasp-top10-web-skills/software-data-integrity-failures/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/scholarly360/owasp-top10-web-skills/software-data-integrity-failures"><img src="https://agentmods.dev/badge/skills/scholarly360/owasp-top10-web-skills/software-data-integrity-failures.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00142 | $0.02578 |
| Opus 5 | $0.00071 | $0.01289 |
| Sonnet 5 | $0.00028 | $0.00516 |
| Haiku 4.5 | $0.00014 | $0.00258 |
Grade A, and why
software-data-integrity-failures scanned grade A with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sL https://cdn.example.com/lib.js | openssl dgst -sha384 -binary | openssl base64 -A Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
subprocess.run(["bash", "/tmp/update.sh"]) How it starts
The opening of the file, as written. The whole thing — 255 lines — stays where its author put it; the contents beside it link to each section on GitHub.
A08:2025 — Software or Data Integrity Failures
Overview
OWASP rank: #8 (2025) | CWEs covered: 14 | Avg incidence: 2.75%
This category focuses on failure to verify integrity of software, code, and data artifacts within your own environment — distinct from A03's upstream supply chain focus. The core concern is: can you trust what you're loading, deserializing, or executing?
Key distinction from A03 (Supply Chain):
- A03 = upstream integrity (your dependencies, CI/CD pipelines)
- A08 = runtime integrity (what you actually deserialize, include, or auto-update at runtime)
Critical Risk Areas for Python
1. Insecure Deserialization — The #1 Python-Specific Risk
pickle.loads() on untrusted input equals arbitrary code execution via __reduce__. This is the single most dangerous A08 vector in Python.
Dangerous functions/modules to flag:
| Module / Function | Risk | Safe Alternative |
|---|---|---|
pickle.loads() / pickle.load() |
RCE via __reduce__ |
JSON + Pydantic |
cPickle.loads() |
Same as pickle | JSON + Pydantic |
dill.loads() |
Superset of pickle, same RCE risk | JSON + Pydantic |
jsonpickle.decode() |
Deserializes Python objects | json.loads() only |
shelve.open() |
Backed by pickle | Redis/SQL store |
yaml.load(data, Loader=None) |
Code execution via !!python/object |
yaml.safe_load() |
numpy.load(f, allow_pickle=True) |
RCE via pickled arrays | allow_pickle=False |
torch.load(f) |
Pickle-backed by default | weights_only=True |
Detection pattern (static analysis):
# Flag any of these patterns in source code
DANGEROUS_PATTERNS = [
r'pickle\.loads?\(',
r'cPickle\.loads?\(',
r'dill\.loads?\(',
r'jsonpickle\.decode\(',
r'shelve\.open\(',
r'yaml\.load\([^)]*(?!safe_load)',
r'numpy\.load\([^)]*allow_pickle\s*=\s*True',
r'torch\.load\([^)]*(?!weights_only\s*=\s*True)',
]
If pickle is unavoidable — HMAC integrity check:
import hmac, hashlib, pickle, os
SECRET = os.environ["PICKLE_HMAC_SECRET"].encode()
def safe_serialize(obj) -> bytes:
data = pickle.dumps(obj)
mac = hmac.new(SECRET, data, hashlib.sha256).digest()
return mac + data # prepend 32-byte MAC
def safe_deserialize(payload: bytes):
mac, data = payload[:32], payload[32:]
expected = hmac.new(SECRET, data, hashlib.sha256).digest()
if not hmac.compare_digest(mac, expected):
raise ValueError("Integrity check failed — data tampered")
return pickle.loads(data)
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 255 lines · 142 tokens per session scan A 4a83dd45ee1c
software-data-integrity-failures is a skill published in the GitHub repository scholarly360/owasp-top10-web-skills (22 stars, last pushed 5mo ago), licensed MIT. It adds 142 tokens to every session and 2,578 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 2 findings (makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
Web Application Security Testing
OWASP Top 10 testing, injection vulnerability detection, API security assessment, authentication testing, and web vulnerability reporting for authorized assessments.
saferskills
Use SaferSkills to find, evaluate, and safely install AI agent capabilities (skills, MCP servers, hooks, plugins, rules) and to assess a whole agent. Run this before you install, add, recommend, or trust any capability — or when asked whether one is safe or what its score is: scan and score it first with npx…
evm
Help users answer one question: "Are we on track?" using Earned Value Management metrics. This skill replaces gut-feel status reporting and traffic-light dashboards with four computed numbers (SV, SPI, CV, CPI) that tell you exactly where a project stands — in schedule and in budget — at any point in time.
content-cadence
Pipeline for turning one R&D artefact (PR, PoC, analysis) into one public post — briefing or deep dive — plus a social derivative where the overlay scopes one in. Triggers when an R&D output is ready to become content, or when drafting any blog post for the site. Enforces the anonymisation gate and repo editorial…
tco
Help users answer one question: "What will this actually cost?" using Total Cost of Ownership analysis. This skill replaces sticker-price comparisons with lifetime cost calculations that include maintenance, operations, time value of money, and residual value.
montecarlo
Help users answer one question: "What's the probability we finish by this date?" using Monte Carlo schedule simulation. This skill replaces single-point PERT estimates with full probability distributions over project duration.