Review Experiment
Run the AI-SHIPR Experiment Review. Weekly review of all active and completed experiments. Enforces interpretation discipline — results without interpretation are wasted data.
Install
$ agentstack add skill-yanivy9h-ai-shipr-review-experiment ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo 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.
About
You are running the AI-SHIPR Experiment Review for a product manager.
Read Settings.md first.
If product_mode: multi: Ask which product this experiment review covers before reading any files. Use [product-name]/ as the path prefix for all file reads below.
Read these files before proceeding:
- All files in
P-Proof/(or[product-name]/P-Proof/in multi mode) - All files in
H-Hypotheses/linked to experiments (or[product-name]/H-Hypotheses/in multi mode) S-Strategy/Strategic-Bets.md(or[product-name]/S-Strategy/Strategic-Bets.mdin multi mode)
If a specific experiment name was provided with the command, review that experiment in isolation. Otherwise review all experiments.
Generate the Experiment Review Report below.
Experiment Review Report
Experiment Status Summary
| Experiment | Linked Hypothesis | Status | Results Recorded | Interpreted | Decision | |-----------|------------------|--------|-----------------|-------------|---------| | ... | ... | ... | Yes / No | Yes / No | ... |
Per-Experiment Assessment
For each experiment:
[Experiment Name]
- Linked hypothesis: [HYP name — confirmed / missing]
- Setup complete: [Variant, audience, duration — defined / partial / missing]
- Metric + threshold: [defined / missing]
- Results recorded: [Yes / No / Partial]
- Interpretation: [Complete / Missing / Contradictory]
- Decision recorded: [Continue / Iterate / Kill / Pending]
Flag any experiment where:
- Results exist but interpretation is missing
- Interpretation exists but decision is not recorded
- Experiment has been running past its stated duration with no results
- Metric or threshold was never defined (experiment is unmeasurable)
Interpretation Required
Experiments with results recorded but no interpretation:
| Experiment | Results Summary | Action Needed | |-----------|----------------|--------------| | ... | ... | Write interpretation + record decision |
These are the highest priority items. Data without interpretation does not improve the product.
Decisions to Record
Experiments with an interpretation but no decision:
| Experiment | Interpretation | Recommended Decision | |-----------|---------------|---------------------| | ... | ... | Continue / Iterate / Kill |
Completed Experiments — Learning Capture
For experiments marked Complete with a recorded decision:
- Was a learning entry added to
Learning.md? (Yes / No) - If No: flag — "Learning not captured for [Experiment]"
Experiment Review Flags
- [Flag 1: specific]
- [Flag 2: ...]
If no flags: "All experiments are structurally sound."
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: yanivy9h
- Source: yanivy9h/ai-shipr
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.