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

Necromancy

skill-dtiger1889-ops-claude-harness-toolbox-necromancy · by dtiger1889-ops

Recover a Claude Code session that died before a CHECKPOINT was written

No reviews yet
0 installs
0 views
view→install

Install

$ agentstack add skill-dtiger1889-ops-claude-harness-toolbox-necromancy

✓ 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-dtiger1889-ops-claude-harness-toolbox-necromancy)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
3d 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 Necromancy? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Recover state from a dead session's on-disk transcript. The script does the heavy lifting deterministically; you NEVER Read raw session JSONL into context (files run to multiple MB -- the digest exists precisely to avoid that).

Script: ~/.claude/skills/necromancy/necromancy.ps1 (run via the PowerShell tool; Windows-first, like the rest of this toolbox's scripts).

  1. LIST. Run with the project folder the user means (default: the folder they're

asking about, usually cwd): & "$env:USERPROFILE\.claude\skills\necromancy\necromancy.ps1" -Project Output: newest-first sessions with title, first ask, size, and checkpoint state ("NO CHECKPOINT WRITE" or "N assistant msgs after last write").

  1. PICK the dead session. The newest file is almost always the CURRENT live session

-- skip it. A dead candidate is: NO CHECKPOINT WRITE, or many assistant messages after the last checkpoint write, matching the timeframe/topic the user describes. If ambiguous, show them the list lines and ask which one.

  1. DIGEST. ... necromancy.ps1 -Project -Session

Sections: FIRST ASK, USER MESSAGES (their decisions -- the gold), FINAL ASSISTANT STATE, FILES WRITTEN/READ, COMMANDS, GIT/COMMITS, VERSIONS, TESTS, ERRORS.

  1. SYNTHESIZE a resume summary FROM THE DIGEST ONLY: what was asked, what got

decided, what was completed vs in flight when it died, and the literal next action. Verify claims about file state against the ACTUAL files (a dead session's last words may describe work it never flushed) before asserting anything as done.

  1. OFFER TO PERSIST. If the project has a CHECKPOINT.md that predates the dead

session, the natural close is updating it (the /checkpoint skill) so the recovered state has a durable home. Do it when the user confirms resuming.

Gotchas:

  • Slug mapping is [^A-Za-z0-9] -> '-' on the project path; the script falls back

case-insensitively (Windows store dirs vary in case; same inode either way).

  • Digest caps per category (precision over recall). If a needed detail is missing,

re-run digest, then at worst grep the single session file for a specific string (grep -o a narrow pattern) -- still never full-Read it.

  • Anchor extraction is regex-only and can misfire; treat ERRORS/COMMITS lines as

leads, not verified facts.

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.