ai.markettrace/agent-feed

The evidence

Every deduction below names the artefact that produced it. Nothing on this page was executed: the source was read, never run, and no running server was contacted.

Code Safety

• mcp_server.py:37mcp-python-path-join-unvalidated (medium/low)

  A filesystem path is assembled from a value that is not a literal and opened without confinement. os.path.join discards the base entirely when the second component is absolute, and neither join nor `/` resolves `..`.

  # from the live server (README §"Refresh the contract"); the version here
  # is what the bridge reports in serverInfo.
  with open(os.path.join(HERE, "tools.json"), encoding="utf-8") as fh:
      _contract = json.load(fh)
  TOOLS = _contract["tools"]

• mcp_server.py:79mcp-python-ssrf-nonliteral-url (medium/low)

  An outbound HTTP request is made to a URL that is not a literal. If any part of it is caller-supplied, the server can be steered at internal hosts — cloud metadata endpoints, admin ports, other tenants.

          method="POST",
      )
      with urllib.request.urlopen(req, timeout=30) as resp:
          body = resp.read().decode("utf-8", "replace")
      if resp.headers.get("Content-Type", "").startswith("text/event-stream"):

Auth Posture

• authentication_model

  no authentication appears in the source we fetched, but this server's network endpoint is a hosted remote, whose authentication can be enforced in front of its code — and no credential header is declared. An undeclared header is not an absent one, so `03` §4's "no authentication" band is not established.

• transport_security

  declared endpoint(s) are HTTPS: https://api.markettrace.ai/mcp; scored 80, not 100: `03` §4 reserves 100 for HTTPS *with HSTS*, and an HSTS header can only be seen by connecting, which `04` §9 does not do. This is an unverifiable band, not an observed absence.

• secret_handling

  no committed credential found, and no environment variable with a credential-shaped name is read anywhere in the source

• authorization_granularity

  no tool registration was recognised anywhere in the source. An MCP server registers tools by definition, so this is a gap in our detection rather than a server with none, and `03` §4's threshold cannot be applied to it. A tool-list handler IS present, so the tools are built in a shape this counter could not enumerate.

Maintenance

• recency

  last commit 10 day(s) before 2026-10-16

• issue_responsiveness

  lifetime median (1 issue(s) in the last 6 months, under `03` §5's floor of 3): 4.8 day(s) to first maintainer comment

• release_cadence

  a release exists but fewer than two fall inside `03` §5's 12-month window, so no median gap can be taken — and §5's remaining bands are all defined on that median. A limit of the methodology, not of this repository.

• bus_factor

  sole maintainer

• repository_signals

  `03` §5 scores this by quartile WITHIN the server's category, which needs a distribution across every scanned server and so cannot be computed while scanning one. It is a post-pass over the crawl and that pass does not exist yet.

Transparency

• license

  LICENSE present with no SPDX-License-Identifier tag; identified as 'MIT' from its text

• readme_quality

  README.md: over the size floor but none of `03` §7's content heuristics matched

• declared_scopes

  specific scope or access documentation is present, but `03` §7's 100 and 60 bands both compare it against scopes INFERRED from static analysis, which is not built yet. Scoring 100 on an empty inference would be a vacuous truth — 'all ∅ inferred scopes are documented'.

• changelog

  no CHANGELOG in the fetched source. `03` §7's partial credit for tagged releases carrying release notes is not applied here — that needs the forge API, which this check does not call

• security_contact

  SECURITY.md present