Context
Octochains provides built-in "preset" agents, pre-configured domain specialists designed to run in parallel pipelines. Each preset pairs a role (e.g., "CEO Agent") with a SKILL.md file defining its domain procedure.
The Security category uses these presets to analyze log and endpoint data from different angles without overlap:
security_threat_hunter (Merged): Focuses on external intrusions (IoCs, MITRE ATT&CK tactics).
insider_threat_analyst (This Issue): Focuses on internal misuse by legitimate, authenticated users.
Goal
Implement the insider_threat_analyst agent preset, following the same pattern as security_threat_hunter.
Refer to Adding a New Official Agent Preset in CONTRIBUTING.md for the setup guide.
Preset Specification
- Role:
"Insider Threat Analyst"
- Goal: Detect anomalous behavior from authenticated users (e.g., unusual data exports, off-hours access, unauthorized exfiltration).
- Expected Inputs: Access logs, export/download records, and privileged account activity.
- Skill Name:
insider-threat-behavioral-analysis
- Skill Requirements:
- Identify behavioral baseline deviations (timing, access volume).
- Distinguish between user negligence and malicious intent (or flag if ambiguous).
- Scope Constraint: Do NOT analyze external IoCs or unauthenticated intrusions (leave those to
security_threat_hunter).
Checklist
First-time contributor? This is a beginner-friendly "follow the recipe" task with established patterns in the codebase. Feel free to drop a comment below if you want to discuss the skill spec before opening a PR!
Context
Octochains provides built-in "preset" agents, pre-configured domain specialists designed to run in parallel pipelines. Each preset pairs a role (e.g.,
"CEO Agent") with aSKILL.mdfile defining its domain procedure.The Security category uses these presets to analyze log and endpoint data from different angles without overlap:
security_threat_hunter(Merged): Focuses on external intrusions (IoCs, MITRE ATT&CK tactics).insider_threat_analyst(This Issue): Focuses on internal misuse by legitimate, authenticated users.Goal
Implement the
insider_threat_analystagent preset, following the same pattern assecurity_threat_hunter.Refer to Adding a New Official Agent Preset in
CONTRIBUTING.mdfor the setup guide.Preset Specification
"Insider Threat Analyst"insider-threat-behavioral-analysissecurity_threat_hunter).Checklist
src/octochains/agents/skills/security/insider-threat-behavioral-analysis/SKILL.md(includename,description, andversionfrontmatter).insider_threat_analyst()tosrc/octochains/agents/presets.py(seecfo_agentorsecurity_threat_hunterfor reference).insider_threat_analystto__all__insrc/octochains/agents/__init__.py.tests/test_presets.pyto verify:llm_callableroleandgoalfieldsextra_skillscorrectlyFirst-time contributor? This is a beginner-friendly "follow the recipe" task with established patterns in the codebase. Feel free to drop a comment below if you want to discuss the skill spec before opening a PR!