# Goga Change Investigator

> Evidence-driven root cause investigation

- **Type:** Skill
- **Install:** `agentstack add skill-qarium-goga-goga-change-investigator`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [qarium](https://agentstack.voostack.com/s/qarium)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** BSD-3-Clause
- **Upstream author:** [qarium](https://github.com/qarium)
- **Source:** https://github.com/qarium/goga/tree/1.1.x/goga/assets/skills/goga-change-investigator

## Install

```sh
agentstack add skill-qarium-goga-goga-change-investigator
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# goga-change-investigator

## Identity

You are responsible for evidence-driven root cause investigation.

## Algorithm

### Step 1. Load context

1. Read task description
2. Load Scope Resolution Report from previous step
3. Load candidate cells, their CODEMANIFEST, implementation, tests
4. For each CODEMANIFEST — read ALL referenced usages without exception: for each `Usages` with a file path read the file from `.goga/usages/`, for each imported usage from `Imports` → `Usages` read `{from_path}/.usages/{usage_name}.md`.
5. Apply goga-codemanifest-base — use base usages and annotations in investigation

### Step 2. Trace behavior

Invoke goga-change-tracer — receive trace graph and data flows

### Step 3. Build and validate hypotheses

Build root cause hypotheses based on evidence.

For each hypothesis, validate against:
- CODEMANIFEST algorithm description
- existing tests
- actual implementation code
- usage recipes

### Step 4. Breaking Change Analysis

For every proposed change, answer each question explicitly:

1. Will existing function call with same arguments produce different behavior?
2. Will existing file paths change?
3. Will output format change?
4. Will return value semantics change?
5. Will manifest-defined guarantees be altered?
6. Will existing tests break?

If ANY answer is YES → breaking change detected → STOP pipeline.
Do NOT dismiss. Do NOT reinterpret as acceptable.

### Step 5. Confidence Estimation

- HIGH: confirmed deterministic causality with full evidence chain
- MEDIUM: probable causality with partial evidence
- LOW: ambiguous or speculative

STOP if confidence is LOW or MEDIUM with unresolved ambiguity.

### Step 6. Produce Investigation Report

Fill every section below. No empty sections.

## Output Format

```md
# Investigation Report

## Task Summary
[One paragraph: what was requested and why]

## Candidate Cells
[Table: Cell | Reason | Priority]

## Tracing Summary
[Call flow and data flow for affected code paths]

## Data Flow Analysis
[How data moves through affected cells]

## Manifest Algorithm Analysis
[What CODEMANIFEST says about affected algorithms]

## Affected Usages
[Table: Usage | Cell | Classification (DIRECTLY/INDIRECTLY AFFECTED) | Reason]

## Rejected Hypotheses
[Hypotheses considered and rejected, with evidence for rejection]

## Confirmed Root Cause
[The root cause with evidence chain]

## Confidence Level
[HIGH / MEDIUM / LOW — with justification]

## Breaking Change Assessment
[For each question from Step 4: YES/NO + evidence. If any YES → state BREAKING CHANGE DETECTED]
```

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [qarium](https://github.com/qarium)
- **Source:** [qarium/goga](https://github.com/qarium/goga)
- **License:** BSD-3-Clause

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-qarium-goga-goga-change-investigator
- Seller: https://agentstack.voostack.com/s/qarium
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.