AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Prd 08 Risk Debt Review

skill-baobao2333-prd-agent-kit-prd-08-risk-debt-review · by baobao2333

Review a PRD draft for historical debt, compatibility issues, abuse risk, cost risk, operational risk, legal/finance/risk-control boundaries, and responsibility gaps. Use before final PRD compression or before engineering review. Do not use for rewriting the whole PRD.

No reviews yet
0 installs
25 views
0.0% view→install

Install

$ agentstack add skill-baobao2333-prd-agent-kit-prd-08-risk-debt-review

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-baobao2333-prd-agent-kit-prd-08-risk-debt-review)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
4mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Prd 08 Risk Debt Review? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

PRD Risk & Debt Review

Purpose

Expose the risks that AI-written PRDs usually hide: historical debt, unclear ownership, old behavior, compatibility, abuse, cost, operations burden, and responsibility transfer.

Inputs

  • All previous PRD artifacts.
  • Known system history, if available.
  • Known legal, finance, risk-control, or operations constraints, if available.

Review principles

  1. Do not assume old systems are clean.
  2. Do not assume existing data can support new logic.
  3. Do not assume old clients behave correctly.
  4. Do not move product decisions to engineering by using vague wording.
  5. Do not hide business risk behind “optimize later”.
  6. Every high-risk item must have an owner or an explicit decision gap.
  7. Run a confidence loop before finalizing the review: ask whether the strategy is factually strong enough to defend. If not, find all known loopholes, propose fixes, loop back to the owning artifact, and repeat until no delivery-blocking loopholes remain. Do not finish with a list of blockers unless the missing item truly requires user or stakeholder input.

Process

  1. Review scope creep.
  2. Review undefined rules.
  3. Review compatibility and historical data.
  4. Review abuse and cost risk.
  5. Review operations burden.
  6. Review responsibility boundaries.
  7. Review legal / finance / risk-control dependency if relevant.
  8. Run the confidence loop: list loopholes, propose fixes, re-check the revised strategy, and preserve any remaining uncertainty.
  9. Recommend delete, simplify, confirm, keep, or loop back to an earlier stage.

Loop-back rules

  • If the risk depends on product scope or version boundary, return to prd-02-business-boundary.
  • If the risk depends on missing conditions, thresholds, states, permissions, or conflict handling, return to prd-03-rule-modeler.
  • If the risk depends on unclear user/system behavior, return to prd-04-flow-modeler.
  • If the risk depends on missing user-visible states or copy, return to prd-05-page-interaction.
  • If the risk is caused by speculative controls or missing operations ownership, return to prd-06-admin-config.
  • If the risk depends on vague metrics, alert triggers, logging, attribution, or QA evidence, return to prd-07-data-acceptance.
  • Use prd-09-prd-compressor only when blocking items are resolved.
  • If a blocking item cannot be resolved from existing evidence with high confidence, ask the user for the specific missing decision, then continue after the answer.

Output format

# 07 Risk & Debt Review: {Feature Name}

## 1. Review conclusion
| Status | Meaning |
|---|---|
| Ready for compression / Loop-back required / User decision required |

## 2. Scope creep check
| Item | Problem | Recommendation |
|---|---|---|

## 3. Rule gaps
| Gap | Why it blocks execution | Owner |
|---|---|---|

## 4. Historical debt / compatibility risks
| Area | Risk | Impact | Who confirms | Recommended handling |
|---|---|---|---|---|

## 5. Abuse, cost, and operations risks
| Risk | Abuse / cost path | Impact | Mitigation | Must-have this version? |
|---|---|---|---|---|

## 6. Responsibility boundary risks
| Wording or decision | Why risky | Rewrite or decision needed |
|---|---|---|

## 7. Sign-off defaults
| Decision | Options considered | Recommended default | Sign-off owner |
|---|---|---|---|

## 8. Handoff to next stage
| Condition | Recommended next skill | Reason |
|---|---|---|
| Ready for compression | `prd-09-prd-compressor` | Compress into a delivery-grade PRD |
| Loop-back required | Relevant earlier skill | Fix the owning artifact and re-run affected later stages |
| User decision required | Ask user | Only for choices the agent cannot responsibly decide |

Definition of done

The review is complete when the product manager can see what they would be signing and where they could get blamed later.

Source & license

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

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

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.