Pipelock Known Limitations: What It Does Not Catch

A standing, public register of what Pipelock does not protect against. Honesty is the moat.

Ready to protect your own setup?

Pipelock publishes what it cannot catch. A security control that claims full coverage is the one to distrust, and a reviewer needs the boundaries drawn to decide what residual risk is acceptable. This page is the standing register. It tracks the documented gaps and, per release, the false-positive and true-positive picture.

The authoritative engineering source is the bypass resistance doc, which lists the evasion techniques Pipelock handles alongside the ones it does not. This page is the operator-facing summary.

Out of scope

These are outside what a network and tool proxy can see. If your threat model includes them, you need an additional control, named in the mitigation column.

LimitationWhyMitigation
Process-level attacksPipelock mediates traffic, it does not isolate the process. An agent that can exec arbitrary processes can open a channel that never touches the proxy.Pair with an OS sandbox (Docker, gVisor, Firecracker) and containment.
Non-HTTP exfiltrationA write to a cloud-synced folder, the clipboard, stdout, or a subprocess that does not use the proxy never reaches Pipelock.Container network isolation so the only egress path is the proxy.
Model compromiseIf the model itself is fine-tuned to be malicious, Pipelock cannot read intent. It sees the network-visible symptoms, not the motive.Defense in depth across the four functions.
Pixel-level steganographyData hidden in image pixels or audio or video sample bits is beyond content inspection. Pipelock strips EXIF, XMP, IPTC, and ICC metadata, so metadata-based steganography is covered, but pixel-level encoding is not.Specialized media analysis if your threat model needs it.
Covert channelsTiming, header ordering, and HTTP/2 frame flags carry bits outside the URL and body content layer. Detecting them needs protocol-level analysis beyond content inspection.Documented gap. Containment limits the reachable channels.

Partial coverage

These are mediated but not absolute. Pipelock reduces the risk without closing it.

LimitationDetailMitigation
Semantic prompt injectionInjection detection is pattern-based, so it is syntactic, not semantic. An injection crafted to match no known pattern can pass.Add custom patterns for your domain. Treat detection as one of four functions, not the whole defense.
Very slow exfiltrationA drip of one byte per hour over days sits below any practical rate limit or entropy threshold.Container isolation prevents it. Without isolation, this is the residual risk.
MisconfigurationAudit mode logs without blocking. An operator who never switches to balanced or strict enforces nothing.Start in balanced mode. Use pipelock audit to generate a tuned config.
IPv6 reachabilityIf the internal CIDR list omits IPv6 ranges, an agent could reach internal services over IPv6.The default config includes the standard IPv6 private and multicast ranges and strips zone IDs before parsing.
Unintercepted CONNECTOn a plain HTTPS CONNECT tunnel without TLS interception, only the destination host is visible, not the path, body, or response.Enable TLS interception for full content coverage on CONNECT traffic.

False-positive and true-positive rates

Detection quality is a moving target, so it gets reported, not asserted. The measurement source is the public agent egress benchmark: a tool-neutral corpus of attack and benign cases scored by an open runner, plus Pipelock’s in-repo evasion test suite for the encoding and Unicode classes.

When a benchmark run is recorded, the measured figures are published here: the benign-traffic false-positive rate and the attack-corpus true-positive (catch) rate, with the corpus version and date. The numbers always come from a real run, never an estimate, and they are not posted on a fixed per-release schedule. Between updates, the benchmark repo holds the current public corpus and runner, so anyone can reproduce the score independently.

No measured snapshot has been recorded here yet.

Where to go next

Frequently asked questions

What does Pipelock not protect against?
Pipelock is a network and tool proxy, not a sandbox. It does not stop process-level attacks that bypass the proxy, non-HTTP exfiltration through channels it does not mediate, a compromised model, pixel-level steganography, or covert channels like timing and HTTP/2 frame flags. It has partial coverage of semantically novel prompt injection and very slow exfiltration. Each gap has a documented mitigation, usually a deployment control such as container network isolation.
Why publish a limitations list at all?
Because a security tool that claims full coverage is the one to distrust. A standing public register lets a reviewer see exactly where the boundaries are and decide what residual risk they accept. It is the same reason Pipelock signs evidence instead of asserting non-bypass: verifiable honesty beats an unverifiable guarantee.
Does Pipelock catch homoglyphs and field-split secrets?
Yes, those are tested and handled, not gaps. The normalization pipeline maps 80-plus confusable characters across major scripts, and cross-request reassembly plus ordered subsequence matching catches secrets split across query parameters and across transports in a session. The real residual gaps are listed below: they are about channels Pipelock does not mediate and attacks that are semantic rather than syntactic.

Ready to protect your own setup?