Repository navigation
docs(threat-model): claim the prompt-injection egress controls - #348
Merged
Merged
Conversation
P-INJECTION-EGRESS: once an answer has read text from outside Lens, outbound tools and setup changes ask first, and chat-room answers keep only the links people, the archive, tools or Lens wrote. Correctness-only, a bypass is VALID-HARDENING (Q26). The Assistant x-prompt-injection row is now claimed; N-PROMPT says marking and flagging make injection less likely and are not claimed; MCP results and the optional classifier are described; Q26 lists #302, #309, #329, #338, #343, #344 and #347. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVez82cREgCJtFk6yHReed
adeelahmad
force-pushed
the
claude/project-thread-mj7ggd
branch
from
October 14, 2026 09:54
685d722 to
59ee829
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Requested by Adeel · project thread
Description
Before:
threat-model.mdlisted the Assistant's x-prompt-injection row as unresolved (Q26), with N-PROMPT disclaiming injection resistance and nothing about what an injection could carry out.After:
VALID-HARDENING(Q26). Voided byai.untrusted_asks: falseandbridge.safe_links: false, with line references. It also says what it doesn't cover: notes and memories, hooks (P-EXT-SCOPE), and an address the read text wrote itself; what the person's browser asks first is P-BROWSER-GUARD's.ai.untrusted_marking,ai.injection_check) make injection less likely and are not claimed.ci/validate_threat_model.pypasses.Last of the prompt-injection series. Rebased on v0.8.1 and regenerated with the same script, so it sits beside the browser and mailbox rows that landed meanwhile.
Motivation and Context
AGENTS.md: new attack surface goes into
threat-model.md. This series added controls, so the model now says which are claimed (egress) and which are only best-effort (marking, flagging), so a security report knows what counts.Types of changes
!in the title)Checklist:
make hookschecks it on commit).ci/validate_threat_model.pychecks prose, YAML and JSON agree)docs/accordingly. (the threat model itself)## UnreleasedinCHANGELOG.md. (n/a)Done questions (AGENTS.md)
ci/validate_threat_model.py; the YAML prose hash refreshed by its tooling.untrusted.defang.🤖 Generated with Claude Code
https://claude.ai/code/session_01BVez82cREgCJtFk6yHReed