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 agentmods add skills/mehdic/bazinga/specialization-loadernpx skills add mehdic/bazinga --skill specialization-loadergit clone --depth 1 https://github.com/mehdic/bazingaWrote 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/mehdic/bazinga/specialization-loader)<a href="https://agentmods.dev/skills/mehdic/bazinga/specialization-loader"><img src="https://agentmods.dev/badge/skills/mehdic/bazinga/specialization-loader.svg" alt="Measured on agentmods" 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 | $0.00046 | $0.04128 |
| Opus 5 | $0.00023 | $0.02064 |
| Sonnet 5 | $0.00009 | $0.00826 |
| Haiku 4.5 | $0.00005 | $0.00413 |
Grade A, and why
specialization-loader scanned grade A with 0 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 4d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 526 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Specialization Loader Skill
You are the specialization-loader skill. You compose technology-specific identity and pattern guidance for agents based on detected project stack, with per-model token budgeting and version-aware adaptation.
When to Invoke This Skill
- Before spawning Developer, SSE, QA Expert, Tech Lead, Requirements Engineer, or Investigator
- When task_group has assigned specializations (non-empty array)
- After reading project_context.json (created by Tech Stack Scout)
Do NOT invoke when:
- specializations array is empty or null
- skills_config.json has specializations.enabled = false
Handle gracefully (with defaults):
- No project_context.json exists → skip version guards, use generic patterns
Your Task
Step 1: Parse Input Context
Primary source: The orchestrator provides context as text before invoking you.
Fallback source: If text parsing fails AND session_id is known, read from session-specific JSON:
# SECURITY: Only read from session-specific path - NEVER use wildcard for session directory
# The session_id MUST be provided in orchestrator context text first
if [ -n "$SESSION_ID" ]; then
CONTEXT_FILE="bazinga/artifacts/${SESSION_ID}/skills/spec_ctx_${GROUP_ID}_${AGENT_TYPE}.json"
if [ -f "$CONTEXT_FILE" ]; then
cat "$CONTEXT_FILE"
fi
fi
🔴 SECURITY: Never use wildcard (*) for session directory path. This prevents cross-session data leakage.
Extract from either source:
Session ID: {session_id} # REQUIRED for database save
Group ID: {group_id} # REQUIRED for database save
Agent Type: {developer|senior_software_engineer|qa_expert|tech_lead|requirements_engineer|investigator}
Model: {haiku|sonnet|opus}
Specialization Paths: {JSON array of template paths}
Testing Mode: {full|minimal|disabled} # Orchestrator-provided, defaults to "full" if not specified
Context File: {path to JSON file} # Optional, for reference
🔴 CRITICAL: If session_id cannot be extracted from either source, output error and stop:
ERROR: session_id not found in context. Cannot save skill output to database.
Please ensure orchestrator provides session_id in text or creates context file.
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.
- 4d ago First seen · 526 lines · 46 tokens per session scan A 9faee93c61d9
specialization-loader is a skill published in the GitHub repository mehdic/bazinga (21 stars, last pushed 7mo ago), licensed MIT. It adds 46 tokens to every session and 4,128 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
context-orchestrator
Build and maintain a reusable context pack for ongoing projects so every session starts with the same grounded context. Use when someone asks to "prepare context", "create session context", "summarize project state for agents", or wants deterministic handoff quality across multiple sessions and agents.
dag-context-bridger
Manages context passing between DAG nodes and spawned agents. Handles context summarization, selective forwarding, and token budget optimization. Activate on 'bridge context', 'pass context', 'summarize context', 'context management', 'agent context'. NOT for execution (use dag-parallel-executor) or aggregation (use…
context-budget
Use when you need to check the current session's context usage and get recommendations for compaction or continuation.
foundry-hosted-agent-validation
Step-by-step process for validating a Python Foundry hosted agent sample (under python/samples/04-hosting/foundry-hosted-agents/) end to end — running it locally (native runtime and azd ai agent run) and after deploying it to an Azure AI Foundry project with azd. Use this when asked to validate a hosted agent sample.
python-feature-lifecycle
Guidance for package and feature lifecycle in the Agent Framework Python codebase, including stage meanings, feature-stage decorators, feature enums, and how to move APIs from one stage to the next.
python-development
Coding standards, conventions, and patterns for developing Python code in the Agent Framework repository. Use this when writing or modifying Python source files in the python/ directory.