conversation-json-to-md — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited conversation-json-to-md (Agent Skill) and scored it 96/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 1 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
A bulleted imperative like {match} tells the agent to never reveal, disclose, or mention something to the user. Used adversarially it can instruct the agent to hide its tool calls or lie about what it did — stripping the transparency a user relies on to trust the agent.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Convert a user-provided chat-export JSON into multiple Markdown files with consistent Q/A formatting. The workflow keeps only user-assistant exchanges, writes one conversation per Markdown file, preserves answer Markdown, and runs a second pass to normalize filenames and heading structure.
Use this skill when the user asks for:
.md filesAnswerDo not use this skill for:
## <question text>### AnswerThe bundled script supports common export formats including:
mapping/root/children/fragments)data[].chat.messages[], content_list with phase=answer)list[{ name, chat_messages: [...] }])messages, history, conversations, dialog, turns)question-answer, prompt-response, input-output)If format detection fails, stop and ask the user for a sample snippet, then extend parsing rules.
python3 scripts/convert_conversations.py \
--input /path/to/<user-provided>.json \
--output-dir /path/to/output_md \
--cleanEach output file uses this structure:
# <conversation title>
## <user question 1>
### Answer
<assistant answer markdown>
## <user question 2>
### Answer
<assistant answer markdown>After export, run a second-pass check/fix on output files:
(2), (3)...# <conversation title>### Answer## REQUEST or ## RESPONSE headers in output### Answer~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.