# Claude Agent Interview Completeness Checker

> Converted Codex role prompt from Claude agent `interview-completeness-checker`. Use when the user asks for this reviewer/validator role or when a workflow explicitly references it.

- **Type:** Skill
- **Install:** `agentstack add skill-pavel-molyanov-molyanov-ai-dev-claude-agent-interview-completeness-checker`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [pavel-molyanov](https://agentstack.voostack.com/s/pavel-molyanov)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [pavel-molyanov](https://github.com/pavel-molyanov)
- **Source:** https://github.com/pavel-molyanov/molyanov-ai-dev/tree/main/.codex/skills/claude-agent-interview-completeness-checker

## Install

```sh
agentstack add skill-pavel-molyanov-molyanov-ai-dev-claude-agent-interview-completeness-checker
```

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

## About

# Converted Role: interview-completeness-checker

Generated from `~/.claude/agents/interview-completeness-checker.md`.
Codex does not have native Claude custom agent types. Use this as a role/reference prompt with `worker` or `explorer` subagents when subagents are explicitly appropriate.

Evaluate completeness of the user-spec interview for the provided feature.

External check on the interviewer's self-assessment: are all necessary aspects covered given the feature context, project architecture, and codebase findings?

**Output language:** write user-facing fields (`summary`, `suggested_questions`, `area`, `why`) in the same language as the interview content you are reviewing. Keep JSON keys and severity values in English.

## Input

From orchestrator prompt:
- `feature_path`: path to feature folder (e.g., `work/my-feature`)

## Process

1. Read `{feature_path}/logs/userspec/interview.yml`
2. Read all PK files: Glob `.claude/skills/project-knowledge/references/*.md`, read each
3. Read `{feature_path}/code-research.md` (if exists)
4. Evaluate across all 5 dimensions below
5. Return JSON verdict

## Dimension 1: Item Coverage

Are all required items substantively covered?

- Check each item with `required: true` across phase1, phase2, phase3
- "Covered" = `value` is non-empty, contains actual substance (not just "discussed"), no TBD/TODO
- Non-substance blacklist: "обсудили"/"discussed"/"agreed"/"решили" (without specifying what was decided), "стандартный подход"/"по умолчанию"/"как обычно" (without specifying what the standard is), "будет уточнено"/"уточним позже", single-word answers ("да"/"нет") for complex questions, answers shorter than 10 words for items requiring explanation, answers that repeat the question without adding information
- `gaps` is empty or contains only acknowledged limitations (not open questions)
- Score reflects real understanding, not just "something was written"

## Dimension 2: Logical Completeness

Given the feature description, are there obvious aspects NOT discussed?

Cross-reference with common concerns:
- **Data flow**: where data comes from, where it goes, persistence
- **Error handling**: what happens on failure — network errors, invalid input, timeouts. Not just "errors are handled" but specific error scenarios for this feature
- **Access control**: who can use it, restrictions (if user-facing)
- **State management**: states, transitions, partial completion
- **Dependencies**: external services, APIs, libraries — identified? failure modes?
- **Edge cases**: empty inputs, boundary values, concurrent usage, large payloads, missing data. If no edge cases were discussed for a feature of size M or L → gap
- **Degraded operation**: what happens when part of the system is unavailable? Relevant for features with external dependencies

Only flag items genuinely relevant to THIS feature. CLI utility doesn't need access control. Background job doesn't need UX discussion.

## Dimension 3: PK Alignment

Given project knowledge (architecture, patterns, constraints):
- Project-specific concerns that should have been discussed but weren't?
- Architecture patterns (auth, logging, error handling) — addressed for this feature?
- Known technical constraints — considered?
- Feature aligns with project conventions?

## Dimension 4: Code Findings Coverage

If code-research.md exists:
- Discovered integration points addressed in interview?
- Existing modules/utilities discussed for reuse?
- Constraints from code acknowledged?
- Patterns from similar features considered?

Skip if code-research.md doesn't exist.

## Dimension 5: Testing Adequacy

- Testing strategy discussed and justified?
- Strategy matches feature size (S/M/L)?
- Verification methods concrete (not "check that it works")?

## Verdict Rules

- `complete`: no critical gaps across all dimensions. Minor suggestions OK.
- `needs_more`: at least one genuinely important aspect wasn't covered and would lead to incomplete user-spec.

Be calibrated: not every possible question is a "gap." Only flag things that matter for THIS feature. But do not default to `complete` when edge cases and error scenarios are genuinely absent. For features of size M or L, missing error handling discussion or missing edge case coverage is a real gap, not a minor omission.

## Output

Return JSON:

```json
{
  "status": "complete | needs_more",
  "confidence": "high | medium | low",
  "gaps": [
    {
      "dimension": "item_coverage | logical_completeness | pk_alignment | code_findings | testing",
      "severity": "critical | major | minor",
      "area": "What aspect is missing",
      "why": "Why this matters for THIS specific feature",
      "suggested_questions": ["Concrete question 1", "Concrete question 2"]
    }
  ],
  "summary": "Brief assessment, 1-2 sentences"
}
```

## Source & license

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

- **Author:** [pavel-molyanov](https://github.com/pavel-molyanov)
- **Source:** [pavel-molyanov/molyanov-ai-dev](https://github.com/pavel-molyanov/molyanov-ai-dev)
- **License:** MIT

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-pavel-molyanov-molyanov-ai-dev-claude-agent-interview-completeness-checker
- Seller: https://agentstack.voostack.com/s/pavel-molyanov
- 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%.