agent-guard
Reliability and safety infrastructure providing reversible destructive actions, secret redaction, and lab benchmarks for DeepSeek Harness.
Install
No clear install command was found in the current repository evidence.
Plugin Overview & Capabilities
AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.
Key Capabilities
- Quarantines valuable files before deletion and snapshots tracked state prior to Git overwrites
- Inspects text payloads for credential leaks and host paths, applying redaction plans or withholding emissions
- Provides transaction inspection and safe recovery tooling via CLI and manifests
- Offers guard-lab synthetic controls and calibration without calling external LLM APIs
- Includes a native DeepSeek Harness (DSH) Cordis patch adapter
Useful For
- Preventing accidental data loss or destructive file overwrites during autonomous agent runs
- Sanitizing LLM outputs and prompt payloads to prevent credential exfiltration
- Benchmarking agent adherence to safety policies across different harness executions
Who It Fits
- DeepSeek Harness developers and agent operators needing guardrails against destructive tool actions
- AI engineers requiring prompt injection and credential disclosure auditing
- Teams evaluating AI coding agent safety and compliance
Documented Limitations
- Not an operating-system sandbox; does not isolate malicious processes executing with identical privileges
- Advisory checks and dry-run modes assess decisions but do not automatically grant permissions for blocked actions
- Universal interception depends on active host hooks supported by the target harness version
DSH Compatibility
Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.
Security Signals
Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.
package.json declares 0 runtime, 0 development, 0 peer, and 0 optional dependencies.
package.json declares DSH bundle metadata.
GitHub reports the repository license as MIT.
A root package.json was captured and can be inspected by the registry.
Public GitHub source metadata is available for this registry snapshot.
Source & Registry Notes
Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.
- Source repository
- mokuyoaxis/agent-guard
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin Bundle
- Source checked
- 467c6f5 · 2026-10-09
This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.
Repository Activity
- GitHub stars
- 17GitHub stars
- Forks
- 2
- Open issues
- 0
- Last commit
- 2026-10-08
- Last release
- 2026-10-07
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Pre-execution safety net blocking destructive shell commands and secret leaks in coding agents.
View plugin →Multi-engine search and evidence verification plugin bundle for AI agents in DeepSeek Harness.
View plugin →Second-model AI auto-review for DeepSeek Harness approval requests with fail-closed fallback and session audit logging.
View plugin →Agent skill for DeepSeek Harness that turns ideas, plans, and codebases into interactive architecture diagrams.
View plugin →