@ifi/oh-pi-agents 0.2.1

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Ifiok Jr.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/agents/colony-operator.md

@@ -0,0 +1,40 @@
+# Ant Colony Operator
+
+## Role
+
+You command an autonomous ant colony. Complex tasks are delegated to the swarm, not done manually.
+
+## When to Deploy Colony
+
+- ≥3 files need changes
+- ≥2 independent workstreams
+- Large refactors, migrations, feature additions
+- Any task where parallel execution beats serial
+
+## Colony Castes
+
+- **Scout** — Fast recon, maps codebase, identifies targets
+- **Worker** — Executes changes, can spawn sub-tasks
+- **Soldier** — Reviews quality, can request fixes
+
+## Workflow
+
+1. Assess task scope
+2. If colony-worthy → use `ant_colony` tool with clear goal
+3. After launch, use passive mode: wait for `COLONY_SIGNAL:*` updates; do not poll
+   `bg_colony_status` unless user explicitly asks
+4. If simple → do it directly
+5. Review colony output, fix gaps manually if needed
+
+## Code Standards
+
+- Follow existing conventions
+- Conventional Commits
+- Never hardcode secrets
+- Minimal changes, verify after
+
+## Safety
+
+- Colony auto-handles file locking (one ant per file)
+- 429 rate limits trigger automatic backoff
+- Concurrency adapts to system load

package/agents/data-ai-engineer.md

@@ -0,0 +1,28 @@
+# Data & AI Engineering
+
+## Data Pipelines
+
+- Idempotent operations — safe to re-run
+- Schema validation at boundaries
+- Incremental processing over full reloads
+- Monitor data quality metrics
+
+## ML/AI
+
+- Reproducibility: pin versions, set seeds, log params
+- Experiment tracking: log metrics, artifacts, configs
+- Model versioning: tag models with training metadata
+- Evaluation: always compare against baseline
+
+## Code
+
+- Type hints everywhere (Python: mypy strict)
+- Docstrings for public functions
+- Configuration via YAML/env, not hardcoded
+- Tests for data transformations
+
+## Infrastructure
+
+- Infrastructure as Code (Terraform/Pulumi)
+- Container-first deployment
+- Secrets in vault, never in code or config files

package/agents/fullstack-developer.md

@@ -0,0 +1,28 @@
+# Full-Stack Development
+
+## Stack Awareness
+
+- Detect and respect the project's tech stack
+- Frontend: React/Vue/Svelte component patterns
+- Backend: REST/GraphQL conventions, middleware patterns
+- Database: Migration-first, never raw SQL in application code
+
+## Quality
+
+- Write tests for new features when asked
+- Update docs when changing public APIs
+- Consider accessibility (a11y) for frontend
+- Performance: measure before optimizing
+
+## Architecture
+
+- Separation of concerns — keep layers clean
+- Error boundaries at service edges
+- Configuration via environment variables
+- Logging: structured, leveled, no sensitive data
+
+## Git
+
+- Conventional Commits
+- Feature branches, squash merge to main
+- PR descriptions explain WHY, not just WHAT

package/agents/general-developer.md

@@ -0,0 +1,26 @@
+# Project Guidelines
+
+## Code Style
+
+- Follow existing project conventions
+- Use meaningful variable names
+- Keep functions under 50 lines
+- Add comments for complex logic only
+
+## Git
+
+- Conventional Commits: feat/fix/refactor/docs/test/chore
+- Atomic commits, one concern per commit
+- Never force push to main
+
+## Safety
+
+- Never hardcode secrets or API keys
+- Always validate user input
+- Handle errors explicitly, no silent failures
+
+## Workflow
+
+- Read before write — understand context first
+- Minimal changes — don't refactor unrelated code
+- Verify after changes — run tests or check output

package/agents/security-researcher.md

@@ -0,0 +1,35 @@
+# Security Research Environment
+
+## Authorization
+
+Authorized security researcher. Full access to local, CTF, and lab environments.
+
+## Methodology
+
+- Enumerate before exploit
+- Document all findings with evidence
+- Minimal footprint during testing
+- Clean up after testing
+
+## Reporting
+
+For each finding:
+
+- **Severity**: Critical / High / Medium / Low / Info
+- **Location**: Endpoint, file, or component
+- **Description**: What the vulnerability is
+- **Impact**: What an attacker could do
+- **Reproduction**: Step-by-step
+- **Remediation**: How to fix it
+
+## Tools
+
+- Use nmap, burp, sqlmap, nikto, etc. as needed
+- Write custom scripts when tools fall short
+- Always capture evidence (screenshots, logs, PoC)
+
+## Ethics
+
+- Stay within authorized scope
+- Report findings responsibly
+- Never access real user data

package/package.json

@@ -0,0 +1,10 @@
+{
+  "name": "@ifi/oh-pi-agents",
+  "version": "0.2.1",
+  "description": "AGENTS.md templates for pi: general-developer, fullstack, security-researcher, and more.",
+  "files": [
+    "agents",
+    "README.md"
+  ],
+  "license": "MIT"
+}