prompt-injection
prompt injection markers
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-08
- Stars
- ★ 408
- Forks
- 19
01 前言 Front Matter
- name
- export-conversation
- description
- Export the current conversation to a markdown file. Use this whenever the user asks to "export", "save", or "write" the conversation to a file, or says something like "export this conversation to X.md".
- disable-model-invocation
- true
02 资源文件 2
- 脚本scripts/find-session-jsonl.sh974 B
- 脚本scripts/parse-session-jsonl.py4.0 KB
03 SKILL.md 正文
Export Conversation
Export the current Claude Code session as a readable markdown file.
Steps
1. Find the session JSONL
Run the bundled script to locate the current session file:
${CLAUDE_SKILL_DIR}/scripts/find-session-jsonl.sh <project-path>
Where <project-path> is the working directory (e.g. /Users/vlad/src/my-project). It prints the path to the newest .jsonl in the project's Claude storage directory — which is the current session.
2. Parse and extract
Run the bundled script:
python3 ${CLAUDE_SKILL_DIR}/scripts/parse-session-jsonl.py <session.jsonl> <output.md> [--from HH:MM:SS] [--until HH:MM:SS]
The script writes # Conversation followed by one section per turn. Each section header shows the role and a human-friendly local time as a subdued subtitle (## user <sub>2:17 pm</sub>). Tool calls that share a record with assistant prose are inlined as brief notes — Bash includes the command, Read/Write/Edit include the file path. Assistant records containing only tool calls are omitted.
Notes
- The JSONL grows during the session. Running this early yields a partial export — that's fine, just note it if relevant.
- The newest
.jsonlis used as a proxy for "current session." This is reliable as long as you run the script promptly; a new session started in another window would produce a newer file. - Agent tool results embedded in the conversation (large code dumps, file contents) should be omitted — they're derivable from the codebase and bloat the export.
04 同仓库更多 Skills
agentsh-policy-edit
Use when adding, removing, or updating rules in an existing AgentSH policy, modifying security permissions, HTTP service declarations, Postgres-family database rules, resource limits, or policy YAML files
agentsh-policy-create
Use when creating a new AgentSH security policy, including agent sandboxes, CI pipelines, development environments, HTTP service gateways, or Postgres-family database access policies
with-allowed
Declares allowed tools
scope-mismatch
declares read-only but writes
oversized
Trips the per-file size limit
minimal
A minimal valid skill