Borrowing it
Nothing to install: this file belongs to ashaychangwani/imprint. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ashaychangwani/imprint/main/.claude/skills/imprint-auth-compile-log/SKILL.mdgit clone --depth 1 https://github.com/ashaychangwani/imprintWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ashaychangwani/imprint/imprint-auth-compile-log)<a href="https://agentmods.dev/skills/ashaychangwani/imprint/imprint-auth-compile-log"><img src="https://agentmods.dev/badge/skills/ashaychangwani/imprint/imprint-auth-compile-log/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ashaychangwani/imprint/imprint-auth-compile-log"><img src="https://agentmods.dev/badge/skills/ashaychangwani/imprint/imprint-auth-compile-log.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00059 | $0.02961 |
| Opus 5 | $0.00030 | $0.01481 |
| Sonnet 5 | $0.00012 | $0.00592 |
| Haiku 4.5 | $0.00006 | $0.00296 |
Grade B, and why
imprint-auth-compile-log scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
T=$(find ~/.claude/projects -name "$SID.jsonl" 2>/dev/null | head -1); echo "transcript: $T" How it starts
The opening of the file, as written. The whole thing — 213 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Investigate a site's auth-tool compile
The authenticate_<site> tool is compiled by a separate agent loop from data tools
(auth-compile-agent.ts → an MCP-server-driven claude-cli/codex session; the live login runs
in auth-verifier.ts on headed cdp-replay). Its logs live in a few places — and the most
important one is not the file you'd guess.
Key gotcha:
~/.imprint/<site>/authenticate_<site>/.compile-log.jsonis hook events only (SessionStart / init / success). The agent's actual tool calls + reasoning are in a separate claude-cli transcript keyed by thesession_idinside that hook log. Step 2 shows how to find it.
Substitute the real site slug for <site> throughout (e.g. amex-fhr, remitly). The auth tool
dir is always ~/.imprint/<site>/authenticate_<site>/.
Step 1 — Ongoing or completed?
SITE=<site>; AUTHDIR=~/.imprint/$SITE/authenticate_$SITE
# Is a teach process still alive? (auth runs HEADED — a visible Chrome window is open mid-run)
pgrep -fl "cli.ts teach $SITE" ; pgrep -afl "Chrome for Testing|Chromium" | head -1
# Completed? these appear only after the auth block finishes
ls -la "$AUTHDIR"/.compile-done.json "$AUTHDIR"/workflow.json "$AUTHDIR"/index.ts 2>/dev/null
| Signal | Meaning |
|---|---|
.compile-log.json mtime advancing, no .compile-done.json |
Ongoing — agent still shaping/verifying |
spinner shows Auth compile: turn N — verify initiate … |
Ongoing, in live verification |
| a headed Chrome window is open + a push/OTP hit your phone | Ongoing, initiate fired — approve it |
.compile-done.json present + workflow.json/index.ts emitted |
Completed (success) — emit() only runs on success |
run log says Auth tool compiled + session stored / compilation failed |
Completed — success / failure |
Step 2 — Find the artifacts (incl. the real transcript)
SITE=<site>; AUTHDIR=~/.imprint/$SITE/authenticate_$SITE
# 1. Hook log — grab the claude-cli session_id from it
SID=$(python3 -c "import json,sys; d=json.load(open('$AUTHDIR/.compile-log.json')); print(next((e.get('session_id') for e in d if e.get('session_id')), ''))")
echo "auth session: $SID"
# 2. The REAL agent transcript (full tool-call conversation), located by session_id
T=$(find ~/.claude/projects -name "$SID.jsonl" 2>/dev/null | head -1); echo "transcript: $T"
# 3. Final outcome
cat "$AUTHDIR/.compile-done.json" 2>/dev/null
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 213 lines · 59 tokens per session scan B 12834b22c9ae
imprint-auth-compile-log is a skill published in the GitHub repository ashaychangwani/imprint (21 stars, last pushed 4d ago), licensed MIT. It adds 59 tokens to every session and 2,961 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
smiles-validation
Strict SMILES validation, structural comparison, and modification verification. Catches invalid LLM-generated molecules.
gini-bug-report
File a locally-captured, already-redacted Gini crash report as a GitHub issue, with the user's consent. Reads the pending crash queue and delegates the actual filing to the github-issues skill.
devharness
Drive and debug a running app via the devharness MCP server - launch or attach to Chrome and Node.js, set breakpoints and logpoints, inspect call stacks and variables, watch console and network, manage dev servers, replay any earlier tool call by its history index, and record reproduction sequences that verify a fix.…
omd-debug
A structured method for finding the root cause of software bugs before changing code. It uses history, reproduction, a narrowly defined scope, and tested explanations.
tc-doctor
Diagnose and fix tinycode environment issues — Ollama install, model health, tool-call capability, directories, agents, skills, connectivity, and container health.
omd-deepen
An architecture review that first finds code areas changed most often in recent commits, then examines those areas for unnecessary complexity, thin modules, and repeated logic. It produces an HTML report of possible improvements without changing files.