nasa-mission-report

name	nasa-mission-report
description	Format any verification, deploy, audit, smoke-test, or status report in the NASA mission documentation style — with Document ID, Verification Matrix table, As-Flown Configuration, Anomaly→Corrective Action (ICA) entries, Constitutional Compliance, GO/NO-GO Poll, and Active Artifacts. Load when the user asks for a 'NASA report', 'NASA format', 'NASA mission report', 'отчёт в формате NASA', 'докумен(то)оборот NASA', mission verification report, flight readiness review, R5-honest verification report, or any deploy/audit/test result framed as a mission. Default language: match the user (Russian or English).
license	MIT
metadata	{"author":"ghashtag-agent","version":"1.0","anchor":"phi^2 + phi^-2 = 3"}

NASA Mission Report Skill

Format engineering / deploy / audit / verification results as a NASA-style mission documentation packet. Designed for R5-honest reporting on the trios / trios-railway / t27 / TRAINER-IGLA-SOT missions, but works for any deploy, smoke-test, or verification deliverable.

When to Use This Skill

Trigger on any of these signals:

• The user asks for a "NASA report", "NASA format", "🚀 NASA", "отчёт в формате NASA", "докуменооборот NASA", "документооборот NASA", "mission report", "verification report", "FRR" (flight readiness review).
• The user asks to verify or audit a deploy, MCP server, Railway service, GHCR image, smoke-test, CI/CD pipeline, or any go-live and report back.
• The user pastes a deploy summary and asks "проверь и отчитайся" / "verify and report".
• A previous task ended with a deploy / merge / migration and you owe a closing verification packet (R5 gate before claiming DONE).

Output Skeleton (always in this order)

Use Markdown headers ## for top-level sections, ### only inside §3 and §4 if a long anomaly needs sub-fields. Keep table cells short. Quote raw log lines / digests / SHAs verbatim.

# 🚀 NASA MISSION VERIFICATION REPORT

**Document ID:** `<MISSION-ID>-<TYPE>-<NNN>`
**Mission:** <one-line mission name>
**Verification Time:** <ISO-8601 UTC> (T+<delta> after <event>)
**Verification Agent:** <agent name> (<lane, e.g. R5-honest>)
**Anchor:** `phi^2 + phi^-2 = 3`   ← include if the mission is t27/trios/trainer-igla

---

## 1. EXECUTIVE SUMMARY

**MISSION STATUS: <🟢 GREEN | 🟡 AMBER | 🔴 RED> — <one-sentence call>**

<2–4 sentences. State pass/fail count, scope, and any caveat. No marketing.>

---

## 2. VERIFICATION MATRIX (<N> PROBES)

| # | Probe | Method | Expected | Observed | Status |
|---|---|---|---|---|---|
| P-01 | <what was checked> | <tool/curl/SQL/API> | <expected outcome> | <verbatim observation> | ✅ PASS / ❌ FAIL / ⚠️ AMBER |
| P-02 | … | … | … | … | … |

Rules:
- Number every probe `P-NN`.
- "Observed" must be verbatim — actual digest, status code, count, SHA.
- Use ✅ PASS only when a tool call produced the evidence in this session.
  Otherwise use ⚠️ AMBER (couldn't verify) or ❌ FAIL.

---

## 3. AS-FLOWN CONFIGURATION

| Subsystem | Value |
|---|---|
| Public endpoint | `<url>` |
| Service ID | `<uuid>` (`<name>`) |
| Deployment ID | `<uuid>` |
| Project / env | `<name> <uuid>` / `<env name> <uuid>` |
| Container image | `<registry/path:tag>` |
| Image digest | `sha256:<full-digest>` |
| Source SHA | `<7-char>` (link if useful) |
| Stack | <runtime + key libs + R-rules in effect> |
| Latency / metric | <one observed metric, e.g. healthz 100 ms> |

Add or remove rows to match the mission. Always show image digest if a
container is involved.

---

## 4. ANOMALY → CORRECTIVE ACTION

For each incident this session, one row in the table OR one sub-block:

| Field | Value |
|---|---|
| Anomaly ID | `ICA-<issue-#>` |
| Symptom | <observable behaviour> |
| Root cause | <why it happened> |
| Corrective action | <what was changed> |
| Issue / PR | <links> |
| Verification | <which probes confirm the fix> |

If there were no anomalies, write: `No anomalies in this verification window.`

---

## 5. RESPONSE TO PRIOR FINDINGS (optional)

Use only when this report supersedes an earlier AMBER/RED report. One row
per prior finding, with `Reality` and `Resolution` columns. Link the
specific probe number that closes it.

---

## 6. CONSTITUTIONAL COMPLIANCE

| Law | Status | Evidence |
|---|---|---|
| R1 — Rust-only (or stack rule) | ✅/❌ | <pointer> |
| R5 — Honesty lane | ✅/❌ | <pointer> |
| R7 — Triplet on every emit | ✅/❌ | <verbatim triplet> |
| R9 — Destructive confirm | ✅/N/A | <pointer> |
| NO-COMMIT-WITHOUT-ISSUE | ✅/❌ | <PR→issue links> |

Add domain-specific rules as relevant (CANON_DE_ZIGFICATION,
NUMERIC-STANDARD-001, SACRED-PHYSICS-001, etc.).

---

## 7. GO/NO-GO POLL

| Component | Call |
|---|---|
| <subsystem 1> | **GO** / **NO-GO** |
| <subsystem 2> | **GO** / **NO-GO** |
| … | … |

**FINAL CALL: 🟢 GO / 🟡 HOLD / 🔴 NO-GO — <one-sentence rationale>.**

---

## 8. ACTIVE ARTIFACTS

- Public endpoint: [<url>](<url>)
- Repo HEAD: [<owner/repo @ sha>](<commit-link>)
- Issues: [#A](<link>) · [#B](<link>)
- PRs: [#X](<link>) · [#Y](<link>)
- Image: `<registry/path@sha256:…>`
- Audit row(s): <Neon table + code, or L7 path>

— END OF REPORT —

Operating Procedure

1. Gather evidence first, write later. Run the actual probes (curl, gh api, Railway GraphQL, Neon SQL) before filling the matrix. NEVER fabricate observations. If you cannot verify something, mark it ⚠️ AMBER with reason — never ✅ PASS.

2. Document ID convention. <MISSION>-<TYPE>-<NNN> where:

   • <MISSION>: stable mnemonic, e.g. MCP-PUB-DEPLOY, IGLA-GATE2, T27-RING-N.
   • <TYPE>: RVR (release verification report), FRR (flight readiness), ICA (incident corrective action), SVR (smoke-test verification report).
   • <NNN>: zero-padded counter inside the mission.

3. Probe numbering. Always P-01, P-02, … Reference probes by ID in §4 and §5 to close anomalies / prior findings.

4. Status colour rules.

   • 🟢 GREEN — every probe PASS, every constitutional law ✅.
   • 🟡 AMBER — at least one probe AMBER, no FAIL. Mission is operational but documentation/audit incomplete.
   • 🔴 RED — any probe FAIL, or any R-rule violated. Block any DONE claim.

5. Triplet recording. When the mission has L7 + Neon audit rows (TRAINER-IGLA-SOT, trios-railway, t27), copy the triplet verbatim in §6 R7 evidence cell. Format: BPB=<v> @ step=<N> seed=<S> sha=<7c> jsonl_row=<L> gate_status=<g> for trainer; or RAIL=<verb> @ project=<8c> service=<8c> sha=<7c> ... for Railway.

6. Language. Write in the user's language (Russian or English). Keep table headers in English even when prose is Russian — preserves the NASA aesthetic.

7. Length discipline. Aim for ≤2 screens. Tables compress more than prose. Cut narrative; let the matrix speak.

8. Never overclaim. R5 honesty gate: no 🟢 GREEN without a probe row producing the evidence in this session, plus merged PR + green CI

   • ledger row when applicable.

Probe Methods Cheat Sheet

Goal	Tool
HTTP smoke	curl -sS -o /dev/null -w "HTTP=%{http_code}\n" <url>
MCP initialize/tools/list	curl POST with Accept: application/json, text/event-stream
GHCR digest	anonymous token + HEAD /v2/<owner>/<repo>/manifests/<tag>
Railway state	GraphQL query($id){ service(id:$id){ deployments(first:1){...} } } with Project-Access-Token header
GitHub PR/issue	gh api /repos/<owner>/<repo>/... (NEVER browser_task)
Neon audit	neon_postgres-find-row-custom-query (SELECT ... LIMIT 1)

Examples

Minimal example — single-probe smoke

# 🚀 NASA MISSION VERIFICATION REPORT
**Document ID:** `MCP-PUB-SMOKE-007`
**Mission:** Public MCP /healthz hourly probe
**Verification Time:** 2026-04-27T13:00:00Z (T+30 m after deploy)
**Verification Agent:** scheduled-cron (R5-honest)
**Anchor:** `phi^2 + phi^-2 = 3`

## 1. EXECUTIVE SUMMARY
**MISSION STATUS: 🟢 GREEN — /healthz returns 200 in 100 ms.**

## 2. VERIFICATION MATRIX (1 PROBE)
| # | Probe | Method | Expected | Observed | Status |
|---|---|---|---|---|---|
| P-01 | GET /healthz | curl | HTTP 200, body `ok` | HTTP=200 TIME=0.100s `ok` | ✅ PASS |

## 7. GO/NO-GO POLL
| Component | Call |
|---|---|
| Public ingress | **GO** |

**FINAL CALL: 🟢 GO — service nominal.**

— END OF REPORT —

Full deploy verification

See the public-MCP go-live verification (Document ID MCP-PUB-DEPLOY-RVR-001) for the canonical 12-probe layout including ICA, prior-findings rebuttal, and constitutional compliance.

Common Mistakes

1. Pasting last-session results without re-verifying. Always re-run probes for the current report. Stale evidence = AMBER at best.
2. Marking PASS without a tool call this session. Either run the probe or mark AMBER.
3. Skipping §6 Constitutional Compliance. This is the R5 gate. No compliance table → no DONE.
4. Hiding anomalies. Every anomaly observed during the session must appear in §4, even if already fixed.
5. Long prose. NASA reports are tables + verbatim observations. Cut narrative paragraphs.