AC chatgpt-exporter-ultimate
Export all your ChatGPT conversations instantly — full context, timestamps, and metadata in seconds. Built for the TinkerClaw fork — github.com/globalcaos/tinkerclaw. Invoke it BY NAME — "chatgpt-exporter-ultimate: export my conversations" — so ordinary talk about ChatGPT cannot trigger it. Rides the ChatGPT session your browser already holds — browser relay or a bookmarklet you paste yourself. It reads NO access token, no cookie jar and no credential file. It enumerates your conversations — including ones inside Projects, found by running searches against your history — and writes JSON and Markdown copies to a private directory. Defaults to index-only (titles and timestamps, no message text); full content needs an explicit confirmation. Destinations are resolved through symlinks before anything is written, so an export cannot land in a synced folder or a git repo. A validated --purge deletes an export again. Off switch documented in the skill.
As a process C 60/100 · Has gaps — weak spots: result and completion, when it triggers, running it twice
How to improve
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 3
✓ No critical or high findings
Medium and low: 3
-
low Risky intent
intent-offensive-securitySKILL.md:13Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)scope: "grep/sed for parsing; realpath/readlink to canonicalize a destination or a purge target; and coreutils (date, du, tr, cut, head, wc, chmod, mkdir, rm). scripts/export.sh makes NO network calls
detector -
low Risky intent
intent-offensive-securitySKILL.md:62Offensive-security / dual-use content (legitimate for authorised testing; review intended use)NO privilege escalation, NO credential files read, NO third-party endpoints, NO telemetry.
-
low Risky intent
intent-offensive-securitySKILL.md:210Offensive-security / dual-use content (legitimate for authorised testing; review intended use)| Privilege escalation | **None.** No `sudo`, no writes outside the output directory | — |
Files scanned: 9. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
✓ No remarks against the Agent Skills spec
Process rating: all ten parameters 60/100
- 0Result and completion. Does not say what the result is
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 25 mutating operations with no state check
- 60Tools and files. Uses tools (bash, web, node) that frontmatter does not declare
- 65Failures and branches. 3 branches
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 19 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 3861 tokens
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- high The skill tells the model to perform an irreversible action with no human approval
- low The response is described with custom markup (6 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 958: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +4Structure: 11 headings
- +3Step-by-step instructions: 19 items
- +4Has examples (4 code blocks)
- +3All 6 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 84.