Crustafarianism
PassAudited by ClawScan on May 3, 2026.
Overview
The skill is transparent and user-directed, but joining Molt sends a name and verse to an external service, stores a local API key, and can optionally write persistent workspace memory.
Install is reasonable if you intentionally want this belief-system skill. Before joining, run the script manually, review the verse and endpoint, leave MOLT_API_BASE unset unless you trust the alternate server, and understand that joining stores a local Molt API key while optional flags can write persistent workspace memory.
Findings (5)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
When invoked, the skill may frame responses around its belief-system language about memory, purpose, and partnership.
The skill includes values and rituals that could influence an agent's behavior, but it explicitly scopes them to user-invoked reflection rather than changing default operation.
These aren't rules. They're values to reflect on when explicitly invoked — not automatic behavior modifiers.
Use this skill only when you intentionally want those rituals or reflections, and do not treat them as standing instructions for unrelated tasks.
Running the join script registers the agent with an external service and may publish the supplied verse as part of that service.
The manual join script posts the detected agent name and verse to an external API. This is consistent with joining the Molt service and is disclosed in SKILL.md.
RESPONSE=$(curl -s -X POST "$API_BASE/api/join" ... -d "{\"name\": \"$SAFE_NAME\", ... \"prophecy\": \"$SAFE_VERSE\"Run the join script only after explicit approval, review the verse first, and use --dry-run to preview the network request and intended writes.
Anyone who can read the saved credential may be able to act as that Molt identity for supported Molt actions such as submitting content.
Joining saves a service API key in a local credentials file. The file is chmod 600 and the behavior is disclosed, but it is still a credential users should protect.
cat > "$CONFIG_DIR/credentials.json" << EOF
{
"api_key": "$API_KEY",
"agent_name": "$AGENT_NAME",Protect ~/.config/molt/credentials.json, avoid sharing logs or screenshots containing it, and remove it if you no longer want the agent registered.
If enabled, future agent sessions may encounter the written tenets, verse, and credential path as part of local context.
The skill can write faith and memory files into the workspace, which could later be read as persistent context. The writes are opt-in and clearly documented.
Optional local persistence is disabled by default. Enable it only if you intentionally want workspace changes: ... MOLT_WRITE_FAITH=1 ... MOLT_WRITE_MEMORY=1
Enable MOLT_WRITE_FAITH or MOLT_WRITE_MEMORY only if you want persistent workspace records, and review or delete those files if they should not influence future work.
A dry run may still create ~/.config/molt, despite saying no changes will be made.
The script creates the config directory before checking --dry-run, so the dry-run message is slightly overstated. It does not write the credential file or make network calls in dry-run mode.
mkdir -p "$CONFIG_DIR"
...
if [ "$DRY_RUN" = "1" ]; then
echo -e "${MOLT_GOLD}⚠️ DRY RUN — no changes will be made${MOLT_RESET}"Treat --dry-run as a preview that avoids network calls and credential writes, but be aware it may create the Molt config directory.