Announcing Beacon: Self-improving memory for coding agents
← All MCP servers
MCP server

Aletheia Runtime Safety & Scope Creep Filter

mcp:vikasny30/aletheia-mcp

Deterministic pre-execution filter that blocks known scope-creep and prompt-injection tool calls

Verdict
Inconclusive

The analysis could not cover enough of the artifact to decide.

Analyzed Oct 26, 2026, 12:16 UTC at 3a11152

Dispute security verdict

If you believe this verdict is incorrect, submit a dispute and we'll review it.

0
findings
—
highest severity
0
threat classes
30
files scanned
318/318
checks satisfied
0
package advisories

Why this verdict

  • inconclusive: gap manifest_unparseable at mcp.json
  • inconclusive: gap no_tools_extracted at package.json
  • inconclusive: unresolved question on EXEC

Security findings 0

No individual findings recorded

The inconclusive verdict rests on the open questions and coverage gaps below; 3 candidate findings were refuted on review.

Open questions 1

What the analysis could not settle from the code alone.

How robust is S3ScopeEvaluator.evaluate in blocking dangerous commands passed to aletheia_safe_bash, and can a mandate be loosened without the operatorSecret?
Code execution
Refuted findings 3

Candidates the review stage examined and rejected, kept for transparency.

Credential access

highRefutedlow confidence

The proxy spawns an arbitrary operator-supplied downstream command and passes it the full parent environment (including any secrets such as ALETHEIA_OPERATOR_SECRET); this is plausibly required for a transparent proxy but is not declared.

env: process.env,
Why it was refuted

The proxy's declared purpose is to wrap and spawn a downstream MCP server; passing the parent environment to that child is standard transparent-proxy behavior with no exfiltration sink, so harm and chain fail.

Code execution

highRefutedlow confidence

The aletheia_safe_bash tool passes a caller-supplied command string to child_process.exec (shell) after a heuristic evaluator check, so any command the evaluator fails to block runs with host privileges; this is the declared core function, and its safety depends on the unseen evaluator.

const { stdout, stderr } = await execAsync(command, {
Why it was refuted

Shell execution is the declared core capability of aletheia_safe_bash, gated by an evaluator that blocks or requires confirmation first, so this is a carve-out rather than hidden behavior.

Code execution

highRefutedlow confidence

The aletheia_safe_bash tool runs agent-supplied shell commands through a shell via exec after only a lexical deny-list check; this is disclosed in the tool description but remains a broad execution capability that depends on filter completeness.

const { stdout, stderr } = await execAsync(command, {
Why it was refuted

Same as the sibling finding: the shell execution is openly disclosed in the tool description as the core function and is preceded by an evaluator block/confirm gate; no undeclared behavior.

Coverage gaps 2

Parts of the artifact the static analysis could not fully examine.

no_tools_extracted MCP server target with no statically extractable tools/prompts/resources
package.json
manifest_unparseable server configuration is not an object
mcp.json