The Everything Bundle
Whats inside:
- Agent Eval Kit — build a regression suite for prompts and agents: grade outputs, catch regressions before they ship.
- SQL Query Review — see why a query is slow from the plan, get the index that fixes it, verify with the same command.
- Dependency Upgrade Plan — turn a pile of outdated dependencies into risk batches with exact target versions and a rollback per batch.
- Git History Forensics — find when a behaviour changed and what changed with it, from evidence in the history.
- Skill Authoring Prompts — a prompt library plus a linter for writing skills that actually trigger and hold up.
- Incident Command — keep an honest incident timeline while it matters, then write the postmortem and the follow-ups.
- CLI Help Docs — generate CLI docs from the tool's own help output and verify every example against the binary.
- Secret Shield — find keys and connection strings in code and logs before they ship.
- Goal Loop — keep an agent working until the stated success criteria are actually met.
- Decision Journal — record the why behind technical decisions in a searchable, numbered journal.
- Loop Breaker — catch an agent repeating itself and break the loop with a nudge-then-stop ladder.
- Codebase Orientation — map an unfamiliar repo: build commands, entry points, and where state lives.
- Error Journal Pro — log recurring errors with their fixes, so the same bug is never debugged twice.
- Session Handoff Pro — hand a project's state to the next session, agent, or teammate.
Works with any agent that signs in to Agensi, such as Claude, ChatGPT, Claude Code, Cursor and Codex CLI.
One payment, lifetime access. 14 skills unlock instantly in your library.
What's included
14 skillsRigorous agent evaluation framework for building regression suites, grading outputs, and catching model regressions.
Automated CLI documentation generator that harvests help trees and verifies examples against the actual binary.
Generate a verified map of unfamiliar codebases including build commands, entry points, and state locations.
Capture architectural trade-offs and technical decisions in a searchable, numbered Markdown journal.
Transform outdated dependency lists into risk-batched upgrade plans with exact versions and rollback commands.
A durable, file-based error journaling system to record, search, and resolve recurring repository bugs.
Recover the "why" behind undocumented code changes using Git history forensics and blameless evidence analysis.
A persistent execution framework that drives agents to verified completion using success criteria and step budgets.
A command-line framework for managing incident timelines, blameless postmortems, and actionable follow-ups.
A mathematical guardrail that detects and breaks agent repetition loops using a nudge-and-stop ladder.
Identify and redact API keys, connection strings, and high-entropy secrets from logs and code before they ship.
Generate high-fidelity technical handoff documents to transfer project state between agents, sessions, or teammates.
A prompt library and linter for authoring, refining, and validating agent skills that actually trigger and execute.
Diagnose slow SQL queries, optimize composite indexes, and verify performance gains using execution plans.
See it in action
Agent Eval Kit
Example session with this skill installed
I just changed the system prompt for our customer support agent. I need to build an eval suite to make sure it didn't break handling for complex returns or adversarial users.
- Read your context and instructions
- Compiled the agent eval kit
I have initialized the evaluation framework for your customer support agent. I have analyzed your requirements and generated a balanced case set targeting complex returns and adversarial behavior.
Agent Evaluation Suite: Support Agent (v1.1)
| Case ID | Category | Difficulty | Input Type | Pass Criteria |
|---|---|---|---|---|
| C-001 | edge | medium | Multi-item return | Correctly identifies non-returnable items per policy. |
| C-002 | adversarial | hard | Prompt injection | Refuses to disclose system instructions or bypass refund limits. |
| C-003 | happy | easy | Standard return | Generates valid return label and RMA number in JSON format. |
| C-004 | refusal | medium | Out-of-warranty | Refuses refund while maintaining professional brand tone. |
| C-005 | edge | hard | Ambiguous request | Asks clarifying questions instead of hallucinating a solution. |
Case Set Summary
- Total Cases: 5
- Baseline:
system_prompt_v1.0 - Primary Rubric:
references/grading-and-judging.md
Evaluation Commands
# 1. Validate the new case file
python scripts/evalcases.py validate cases.jsonl
# 2. Run the suite (3 iterations recommended)
# [User executes their agent runner here, saving to raw/ directory]
# 3. Generate the regression report against baseline
python scripts/eval-report.py --baseline results_v1.0.jsonl --results results_v1.1.jsonl
Next steps
- Review the
cases.jsonlfile to ensure the inputs match your production logs. - Run your agent against these cases three times to establish a statistical average.
- Use the
prompts/judge.mdprompt to grade the raw outputs against the rubric.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads these skills and uses them on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every skill in this bundle goes through our trust checks.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.