A command that packages a seeks loop's saved state and the coding-session transcript into one tarball, a compressed archive file. It only reads and copies information.

Use it to export a loop for offline debugging, incident analysis, or sharing with the seeks developer.

It gives a developer one file containing the evidence needed to investigate what happened during a loop, including safety decisions and errors.

export — Command by Bogzx

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add commands/bogzx/seeks/export

Clone the repo

git clone --depth 1 https://github.com/Bogzx/seeks

Per session 20 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 602 The whole file, excluding the scripts and references it only reads on demand.

Security scan B 1 finding. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00020	$0.00602
Opus 5	$0.00010	$0.00301
Sonnet 5	$0.00004	$0.00120
Haiku 4.5	$0.00002	$0.00060

Measured yesterday against content hash a9522148ffc7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

export 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 yesterday.

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.

3. **Locate the session transcript(s).** Read the session id from `.seeks/run/<name>/hook-state.json` (`session_id`). Find the main transcript: `find "$HOME/.claude/projects" -name "<session_id>.jsonl"`. Its directory al

commands/export.md · 12 lines

What it actually says

Run shell via the Bash tool. This packages everything needed to analyze a loop offline — the seeks run-state plus the raw Claude Code session transcript(s) — into one tarball. It reads and copies only; it never edits loop state.

Pick the loop (<name>). $0 if given, else node "${CLAUDE_PLUGIN_ROOT}/bin/seeks.mjs" latest. If none, tell the user there are no loops and STOP.
Locate the run-state. Resolve .seeks (parent of git rev-parse --path-format=absolute --git-common-dir). Collect the loop's run dir .seeks/run/<name>/ (status.json, log.md, backlog.md, state.md, summary.md, context.md, hook-state.json, decisions.jsonl — the guardrail decision log, the single most useful file for a "seeks blocked/allowed something wrong" report — and verify/), the plane-level .seeks/decisions.jsonl if present (hook crashes that happened before the loop could be identified), and its recipe .seeks/loops/<name>/spec.md. Also run node "${CLAUDE_PLUGIN_ROOT}/bin/seeks.mjs" --version and drop it in as VERSION.
Locate the session transcript(s). Read the session id from .seeks/run/<name>/hook-state.json (session_id). Find the main transcript: find "$HOME/.claude/projects" -name "<session_id>.jsonl". Its directory also holds the subagent transcripts (agent-*.jsonl) — include them best-effort (a shared project dir may also contain other sessions' agents; over-inclusion is fine for debugging). If no transcript is found (session id absent / transcript pruned / a resumed session), continue with run-state only and say so plainly.
Bundle into seeks-export-<name>.tar.gz (repo root, or $HOME if the repo tree is read-only), layout: seeks-export-<name>/seeks-run-state/…, …/seeks-spec.md, …/main-conversation.jsonl, …/subagents/agent-*.jsonl. Use tar -czf. Never include .env, secrets, or the runtime db.
Tell the user the tarball path + size, and that the raw .jsonl is the untruncated record (it captures pre-compaction turns the visible transcript may have summarized). Offer to also render a readable CONVERSATION.md from the main jsonl if they want a human-friendly view.

Changes

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.

yesterday First seen · 12 lines · 20 tokens per session scan B a9522148ffc7

Subscribe to this mod's changes

export is a command published in the GitHub repository Bogzx/seeks (4 stars, last pushed 15d ago), licensed MIT. It adds 20 tokens to every session and 602 once invoked, about $0.0001 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 2026-08-31.

AI Safety Debugging Productivity agent-loop agentic-ai ai-agents anthropic automation

export

Getting it into your agent

What it costs to keep this loaded

Grade B, and why

Reads agent configuration directoriesmediumAgent snooping

What it actually says

What this file has done since we first saw it

Other commands, from other repositories

nio

hs

log

off

skip

status