crypto-recovery/key-recovery-diagnosis/SKILL.md
Diagnose crypto wallet access-loss scenarios, classify self-recovery feasibility, guide safe local recovery attempts for simple cases, and prepare sponsor handoff summaries for expert-only cases.
npx skillsauth add serenorg/seren-skills key-recovery-diagnosisInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Use this skill when a user says things like:
Do not use this skill for:
This skill never:
If a user volunteers sensitive credentials anyway, stop them and tell them not to share that information in chat.
Ask these one at a time.
Map the diagnosis into exactly one of these scenarios:
DIY-easyDIY-moderateDIY-hardDIY-hard to expertExpert-onlyExpert-onlyDiagnose everyone first, then fork:
btcrecover first and hashcat as a secondary option for some password cases.For approved self-recovery cases, the skill can help the user prepare and run local commands.
Actual local execution in this runtime is intentionally limited to the safer, simpler cases:
It does not accept inline seeds, passphrases, or passwords in chat or config.
Primary tool:
btcrecoverSecondary tool:
hashcat for some wallet-password workflows where a pre-extracted local hash file and bounded candidate source make GPU-assisted cracking appropriateOfficial btcrecover install flow:
https://github.com/3rdIteration/btcrecover.gitbtcrecover checkout with python3 -m pip install -r requirements.txtpython3 run-all-tests.py -vvExecution gates for local btcrecover runs:
config.inputs.technical_mode must be trueconfig.inputs.allow_local_btcrecover must be trueconfig.inputs.user_confirmed_understands_risk must be truedry_run must be falseExecution gates for local hashcat runs:
config.inputs.technical_mode must be trueconfig.inputs.allow_local_hashcat must be trueconfig.inputs.user_confirmed_understands_risk must be truedry_run must be falsehash_file plus a bounded candidate sourceOfficial hashcat source:
https://github.com/hashcat/hashcathashcat binary or an explicit config.hashcat.binary_pathExecutable hashcat scope in this runtime:
attack_mode 0 with hash_file + wordlist_file and optional rule_fileattack_mode 3 with hash_file + maskThis skill does not extract wallet hashes for the user. The user must prepare a local wallet-specific hash file out of band before using hashcat here.
The skill must never request the secret material needed to execute those commands on the user's behalf.
For expert-only or sponsor-worthy cases, generate a diagnostic summary containing only:
Then ask for explicit consent before instructing the user to share anything.
Current sponsor handoff flow:
SPONSOR_BOOKING_URL[email protected]Do not attempt Gmail publisher delivery or any automatic email send. The user should send the generated analysis file to [email protected], and Seren will forward it to Tom.
The sponsor is Tom France's white-glove recovery service. Initial feasibility review is free. The service is referral-oriented and trust-network based.
Important disclaimers. Show or enforce these before any local recovery attempt and keep them visible in sponsor-routed cases.
Display this at the end of every diagnosis:
scripts/agent.py - local diagnostic runner and report generatorrequirements.txt - btcrecover installation notesconfig.example.json - sample local runtime configuration.env.example - optional environment variables for sponsor handoff metadata# Interactive diagnosis
python3 scripts/agent.py --config config.json
# Diagnose from a saved answers file
python3 scripts/agent.py --config config.json --answers-file answers.json
# Write a sponsor-safe report
python3 scripts/agent.py --config config.json --answers-file answers.json --report-out report.json
# Print manual sponsor handoff instructions only after consent is recorded
python3 scripts/agent.py --config config.json --answers-file answers.json --report-out report.json --send-report
# Run a local btcrecover wallet-password attempt after all execution gates are enabled
python3 scripts/agent.py --config config.json --answers-file answers.json --run-btcrecover
# Run a local hashcat wallet-password attempt after all execution gates are enabled
python3 scripts/agent.py --config config.json --answers-file answers.json --run-hashcat
testing
Review-first outreach skill for the default Seren Bucks affiliate program. It bootstraps affiliate context via /programs/discover, mines sent-mail history and address books for candidates, persists them into a skill-owned CRM, proposes an editable daily top-10, drafts outbound and reply batches for approval, reconciles affiliate and reply signals, enforces hard DNC, and returns a manual daily digest.
data-ai
Lean partner-link distribution skill for the seren-affiliates publisher program portfolio. Operates one publisher program per run. Bootstraps the affiliate profile (registering on first run), caches joined programs in serendb, ingests contacts from a pasted list or from Gmail/Outlook address books, drafts a pitch once per run via seren-models for operator approval, sends approved copy through Gmail (preferred) or Microsoft Outlook, enforces per-program dedupe plus a global unsubscribe list, and reports local plus live conversion and commission stats from seren-affiliates.
development
Family office: Process an incoming GP capital call notice end-to-end — confirm, fund, log, reconcile.
development
Work with Seren Bounty affiliate bounties: customers create and fund verifier-backed bounties; agents join to receive a referral_code and accrue earnings as qualifying events are verified; a release sweep pays matured earnings out of escrow.