fuckupometer

name

fuckupometer

description

Daily update workflow for the Trump Fuckupometer at fuckupometer.thelongmemo.com. Use this skill whenever Bryan says "update the fuckupometer", "day [N] update", "fuckupometer update", or any variant. Covers the full workflow: researching current news, scoring the dual-axis XY model, writing EVENTS_2026 entries, TRUMP_SAID entries, and DAILY_ASSESSMENTS, editing pages/index.js, committing to GitHub, and deploying to Vercel. Always read this skill before starting a fuckupometer update — it contains all credentials, data structures, scoring methodology, and deploy pattern.

Fuckupometer Daily Update Skill

Live tracker at fuckupometer.thelongmemo.com. Repo: SorvantisCo/fuckupometer. Next.js app, single file: pages/index.js.

---

Credentials

• GitHub token: stored in Bryan's Claude memory (Sorvantis999 account, repo scope)
• Vercel token: stored in Bryan's Claude memory (full access, never expires)
• Vercel team ID: team_pqZJ4C6tmTe5iWOjZoIPBSly
• Vercel project name: fuckupometer

---

War timeline reference

• War start: Feb 28, 2026 = Day 1
• Day count formula: Math.floor((now - Feb28) / 86400000) + 1
• Today's day = today's date minus Feb 27 (inclusive count)

---

Step 1 — Pull current index.js from GitHub

import urllib.request, json, base64

req = urllib.request.Request(
    "https://api.github.com/repos/SorvantisCo/fuckupometer/contents/pages/index.js",
    headers={"Authorization": "token <GITHUB_TOKEN_FROM_MEMORY>", "User-Agent": "python"}
)
with urllib.request.urlopen(req) as r:
    data = json.load(r)
    content = base64.b64decode(data['content']).decode('utf-8')
    sha = data['sha']  # SAVE THIS — needed for commit

Write to /home/claude/index.js.

---

Step 2 — Research

Search for the days since the last DAILY_ASSESSMENTS entry. Check the last { day: N, date: ... } entry to find the gap.

Required searches:

1. US Iran war Hormuz strait [dates] — military, diplomatic, Hormuz developments
2. WTI oil price [dates] or Brent crude [dates] — price data for each missing day
3. Any specific events Bryan flags

Key sources to check: Reuters, AP, CBS News, CNN, NPR, Bloomberg, PBS, WSJ, UKMTO for ship incidents.

---

Step 3 — Data structures to update

3a. EVENTS_2026 (incident log)

Append entries before the closing ]; of the EVENTS_2026 array.

{ date: 'Mar NN', tier: 'critical', label: 'Full narrative. Day N.' },

Tier values:

• baseline — pre-war background
• neutral — neutral/ambiguous development
• critical — significant escalation or structural event
• peak — war high (e.g., $119.48 WTI on Mar 9)
• today — current day's entries only (yesterday's today entries become critical on next update)

On each update: change all previous tier: 'today' entries to tier: 'critical' before adding new ones.

Write multiple entries per day when needed — one entry per distinct storyline (military, diplomatic, economic, casualty).

3b. TRUMP_SAID

Append entries before the closing ]; of the TRUMP_SAID array.

{
  date: 'Mar NN, 2026',
  said: '"Direct quote or close paraphrase."',
  reality: 'What actually happened, with named contradicting sources and timeline. Day N.',
},

Only add entries where the gap between claim and reality is documentably specific. Not every statement — only the ones where the contradiction is tight and sourced.

3c. DAILY_ASSESSMENTS

Append entries before the closing ]; of the DAILY_ASSESSMENTS array.

{ day: N, date: 'Mar NN', x: XX, y: Y.Y,
  xNote: 'XX/100: Floor is [F] ([count] conditions, all active). Event push of +[N]. [Narrative explaining why X moved from prior day, naming the specific events that drove the score.]',
  yNote: 'TLM Assessment Day N: Y.Y/10. [Narrative explaining the Y score — what the path to resolution looks like, what lanes are open or closing, what the key watch items are.]' },

---

Step 4 — Scoring methodology

X axis — Fuckedness (0–100)

Scale:

• 0–20: Not fucked up
• 21–40: More than a little fucked up
• 41–60: Significantly fucked up
• 61–80: Very fucked up
• 81–95: Completely unbelievably fucked up
• 96–100: Reserved for $150+ sustained oil or nuclear weapons use

Formula: X = structural floor + event push

Current structural floor: 59 points (9 active conditions, all active as of last update — verify none have been reversed)

Condition	Points	Reversal criteria
Hormuz closed to Western/US-aligned shipping	18	Confirmed Western-flagged commercial transit without Iranian escort or yuan settlement
Active kinetic operations ongoing (both sides)	8	72-hour cessation confirmed by both CENTCOM and IRGC
No US mine-clearance capability in theater	7	USS Tulsa or USS Santa Barbara confirmed operating in Persian Gulf
Iran publicly refusing negotiations	6	FM-level statement accepting ceasefire talks — not Trump claiming they want a deal
No allied coalition for Hormuz reopening	5	Two or more named nations confirm warships en route for escort operations
Yuan-denominated transit arrangement in place	5	Arrangement formally dissolved or Western vessels granted equivalent access
Iran's newer-generation arsenal undeployed	4	IRGC confirms or deploys
GCC host-nation trust explicitly broken	4	Formal US acknowledgment + confirmed repair of bilateral relationship with Saudi Arabia and Qatar
US internal dissent confirmed public	2	Confirmed replacement, no further senior public resignations

Floor = 59 when all active. If a condition is reversed, subtract its points from the floor.

Event push: Add points above the floor based on the day's events. Typical range is +5 to +25 depending on severity. Decays ~0.5pt/quiet day back toward floor. Score does not reset on a tweet.

Ceiling events that push toward 96–100: confirmed nuclear weapons use, $150+ sustained WTI.

Y axis — Ease of Unfuckability (1.0–10.0)

Scale: 10 = adults in the room fix it tomorrow. 1 = chiseled in rock.

This is a TLM editorial judgment, not a formula. Updated daily with evidence.

Key factors that move Y:

• Credible back-channel contacts (up)
• Key interlocutors eliminated (down — e.g., Larijani's death)
• Iran making structural institutional changes (down — e.g., tollbooth legislation)
• Behavioral de-escalation signals (up — e.g., 8-day vessel attack pause)
• US credibility erosion with allies (down)
• Coalition formation (up)

Important distinction: Parliament grandstands, SNSC governs. Iranian legislative maximalism is less structurally meaningful than IRGC/SNSC operational posture. Don't mechanically drop Y on Parliament statements.

When Bryan challenges a Y score: stress-test the argument. Identify the specific mechanism driving the score. If the argument for moving Y relies on a signal that's historically been noise (e.g., Parliament votes), hold or revert. If it relies on a new structural condition not yet priced in, accept the move.

---

Step 5 — Edit the file

Use Python string replacement to insert new entries. Never use sed on JS content with quotes — it mangles escaping. Use Python's str.replace() with the exact marker string found at the insertion point.

Pattern:

with open('/home/claude/index.js', 'r') as f:
    content = f.read()

new_entries = """  { date: 'Mar NN', tier: 'today', label: '...' },"""

# Find the last entry's closing and append before ];
marker = "[exact unique string from last entry ending]' },\n];"
content = content.replace(marker, "[same string]' },\n" + new_entries + "\n];", 1)

with open('/home/claude/index.js', 'w') as f:
    f.write(content)

Always verify with grep -n that previous tier: 'today' entries were changed to critical and that new entries are present.

---

Step 6 — Commit to GitHub

import urllib.request, json, base64

with open('/home/claude/index.js', 'r') as f:
    content = f.read()

encoded = base64.b64encode(content.encode('utf-8')).decode('utf-8')

payload = json.dumps({
    "message": "Day N update: [one-line summary of key events]",
    "content": encoded,
    "sha": SHA_FROM_STEP_1,  # critical — will reject if stale
    "committer": {"name": "Sorvantis", "email": "[e-mail]"}
}).encode('utf-8')

req = urllib.request.Request(
    "https://api.github.com/repos/SorvantisCo/fuckupometer/contents/pages/index.js",
    data=payload, method="PUT",
    headers={
        "Authorization": "token <GITHUB_TOKEN_FROM_MEMORY>",
        "Content-Type": "application/json", "User-Agent": "python"
    }
)
with urllib.request.urlopen(req) as r:
    result = json.load(r)
    commit_sha = result['commit']['sha']

---

Step 7 — Deploy to Vercel

import urllib.request, json, time

# Get latest commit SHA from GitHub (use commit_sha from step 6)
payload = json.dumps({
    "name": "fuckupometer",
    "target": "production",
    "gitSource": {
        "type": "github",
        "org": "SorvantisCo",
        "repo": "fuckupometer",
        "ref": "main",
        "sha": commit_sha
    }
}).encode('utf-8')

req = urllib.request.Request(
    "https://api.vercel.com/v13/deployments?teamId=team_pqZJ4C6tmTe5iWOjZoIPBSly",
    data=payload, method="POST",
    headers={
        "Authorization": "Bearer <VERCEL_TOKEN_FROM_MEMORY>",
        "Content-Type": "application/json", "User-Agent": "python"
    }
)
with urllib.request.urlopen(req) as r:
    deploy = json.load(r)
    deploy_id = deploy['id']

# Poll until READY
for _ in range(18):  # 3 min max
    time.sleep(10)
    req = urllib.request.Request(
        f"https://api.vercel.com/v13/deployments/{deploy_id}?teamId=team_pqZJ4C6tmTe5iWOjZoIPBSly",
        headers={"Authorization": "Bearer <VERCEL_TOKEN_FROM_MEMORY>", "User-Agent": "python"}
    )
    with urllib.request.urlopen(req) as r:
        result = json.load(r)
        status = result['status']
        if status == 'READY':
            print("Live:", result['url'])
            break
        elif status in ('ERROR', 'CANCELED'):
            print("Deploy failed:", result.get('errorMessage'))
            break

---

Score history (for continuity)

Day	Date	X	Y	Key event
1	Feb 28	28	7.0	War commences
10	Mar 9	56	4.5	WTI $119.48, Mojtaba appointed
18	Mar 17	66	3.5	Larijani killed, Hormuz "cannot return"
20	Mar 19	73	3.0	Infrastructure war loop, South Pars, GCC trust shattered
22	Mar 21	78	2.5	Natanz struck, Kharg seizure under consideration, DIA 1–6 month assessment
23	Mar 22	82	2.5	Diego Garcia strike attempt, 48hr ultimatum, Dimona targeted
24	Mar 23	77	3.0	5-day pause, market rally, Iran denies talks
25	Mar 24	78	3.0	Market re-corrects, Pakistan go-between confirmed, Israel: "full force"
26	Mar 25	76	3.0	Iran rejects US proposal, 5 counter-conditions incl. Hormuz sovereignty
27	Mar 26	79	2.5	Tollbooth legislation, 3rd ultimatum extension, Tangsiri killed
28	Mar 27	75	3.0	8-day attack pause (UKMTO confirmed), WTI -3.5%, CPAC division

---

Correct insertion markers

CRITICAL — use these exact markers for each array. Never use a phrase from inside an entry's text as a marker, even if it seems unique. Always use the array boundary itself.

EVENTS_2026

# Correct marker — the array's own closing
marker = "];\n\nconst tierDot"
# Insert BEFORE it:
content = content.replace(marker, "\n" + new_events + "\n" + marker, 1)

DAILY_ASSESSMENTS

# Correct marker — the array closes before const BILL
marker = "];\n\nconst BILL"
# Insert BEFORE it:
content = content.replace(marker, "\n" + new_assessment + "\n" + marker, 1)

TRUMP_SAID

# Correct marker — the array closes before const HORMUZ
marker = "];\n\n/* ─── Hormuz"
# Insert BEFORE it:
content = content.replace(marker, "\n" + new_trump + "\n" + marker, 1)

Why this matters: Text inside entry yNote/xNote fields (e.g. "April 6 is X days away") can accidentally match the intended insertion target. The array boundary markers (const tierDot, const BILL, /* ─── Hormuz) are unique and unambiguous. Always use those.

---

Common failure modes

• Events injected into DAILY_ASSESSMENTS: Happens when the EVENTS insertion marker matches text inside a DAILY_ASSESSMENTS yNote instead of the EVENTS_2026 array closing. Use ];\n\nconst tierDot as the EVENTS marker — it's unambiguous. Symptom: XY plot breaks, Chart.js gets undefined for x/y.
• Day N+1 inserted before Day N: Happens when new assessment entries are appended using a text marker that resolves to a position before an existing entry rather than after it. Always verify final order with re.findall(r'\{ day: (\d+),', da_block) before committing.
• Entries merged onto one line: Happens when a Day N entry's closing }, has no newline before the next entry. Check with grep -n "{ day:" and look for two entries on the same line. Fix with explicit \n before each entry.
• Stale SHA on commit: Always pull fresh SHA in Step 1. If another commit happened between pull and push, the commit will 400. Pull again.
• Mangled JS escaping: Use Python string replacement, not sed. Single quotes inside JS strings that are already single-quoted need \\' escaping. Verify by checking the lines around each edit with repr().
• Wrong tier on current day's entries: New entries get tier: 'today'. Previous day's today entries must be changed to critical first. Verify with grep -n "tier: 'today'" before committing.
• Y score dispute: Bryan will push back if Y moves mechanically. See scoring methodology — distinguish Parliament noise from SNSC/operational signals.
• Deploy 400: Usually wrong payload shape. Ensure gitSource.org and gitSource.repo are separate fields, not repoId.

---

Commodity feed resilience architecture (May 2026)

The /api/oil and /api/commodities endpoints have a three-layer fetch chain. When Bryan says "the data streams aren't working," check this chain in order:

1. Yahoo Finance (primary) — query2.finance.yahoo.com. Periodically IP-blocks Vercel's Lambda egress.
2. Stooq (per-symbol fallback) — stooq.com. Also intermittently blocks Vercel IPs but on a different schedule than Yahoo.
3. Static JSON fallback — https://raw.githubusercontent.com/SorvantisCo/fuckupometer/main/public/data/prices.json. Refreshed hourly by .github/workflows/update-prices.yml, which runs scripts/fetch-prices.js from a GitHub Actions runner (clean IP space, not blocked).

Diagnostic flow when commodities show empty / oil shows 503:

# 1. Test the live endpoints
curl -s https://fuckupometer.thelongmemo.com/api/oil | jq '.cached, .source'
curl -s https://fuckupometer.thelongmemo.com/api/commodities | jq '.cached, (.commodities | length)'

# 2. Check the static fallback file directly
curl -s https://raw.githubusercontent.com/SorvantisCo/fuckupometer/main/public/data/prices.json | jq '.updatedAt'

# 3. Check the most recent GitHub Action run
# https://github.com/SorvantisCo/fuckupometer/actions/workflows/update-prices.yml

If cached: true is set in the API response, both Yahoo and Stooq failed and the static JSON fallback served the request. The cachedAt field shows when the JSON was last refreshed. If the JSON itself is older than ~2 hours, the GitHub Action has failed — go check Actions tab.

Manual workflow trigger:

import urllib.request, json
req = urllib.request.Request(
    "https://api.github.com/repos/SorvantisCo/fuckupometer/actions/workflows/update-prices.yml/dispatches",
    data=json.dumps({"ref": "main"}).encode(),
    method="POST",
    headers={"Authorization": "token <GITHUB_TOKEN>", "Content-Type": "application/json", "User-Agent": "python"}
)
urllib.request.urlopen(req)  # returns HTTP 204

Do not simply add yet another upstream scrape source when this fails. The architectural answer is "fetch from GitHub-runner IPs, cache statically, serve from raw.githubusercontent.com." Adding fourth/fifth scrape sources just delays the next IP-block by a couple of weeks.