self-observability

A JSONL activity log of an agent's own operations. JSONL is a text format with one JSON record per line, and the log records actions such as file reads, commands, writes, timing, and failures.

In plain words
What is it for?
Use it to review an agent's activity, measure time spent on a connected path such as Slack, or find the latest failed command.
Why use it?
It makes the agent's recent work inspectable instead of leaving activity invisible. You can search the log to check what it accessed, how long operations took, or which command failed.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/nexi-lab/nexus/self-observability
Clone the repo
git clone --depth 1 https://github.com/nexi-lab/nexus
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 743 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00743
Opus 5 $0.00000 $0.00371
Sonnet 5 $0.00000 $0.00149
Haiku 4.5 $0.00000 $0.00074

Measured 3d ago against content hash ffabfdf6b81d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

self-observability scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

docs/agents/self-observability.md · 80 lines

How it starts

The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Self-Observability

Every Nexus agent has access to a JSONL log of its own activity at:

/.activity/{utc_date}/{agent_id}.jsonl

Use the same cat, grep, and jq you already use to inspect any other mount.

Schema

{"ts":"2026-05-09T23:42:11.043Z","kind":"op","op":"read","path":"/s3/bucket/foo.txt","bytes":12834,"ms":43}
{"ts":"2026-05-09T23:42:12.110Z","kind":"exec","cmd":"grep needle /gh/owner/repo/README.md","exit_code":0,"ms":215}
{"ts":"2026-05-09T23:42:14.221Z","kind":"op","op":"write","path":"/local/notes.md","bytes":412,"ms":8}

ts is ISO-8601 UTC with millisecond precision and a Z suffix. cmd is truncated to 4 KB; truncated records carry "cmd_truncated": true.

Examples

What did I read in the last hour?

grep '"kind":"op"' /.activity/2026-05-09/me.jsonl | grep '"op":"read"' | tail

How much time did I spend on Slack today?

jq 'select(.path | startswith("/slack/")) | .ms' /.activity/2026-05-09/me.jsonl \
  | awk '{s+=$1} END {print s}'

What was my last failed command?

grep '"kind":"exec"' /.activity/2026-05-09/me.jsonl \
  | jq 'select(.exit_code != 0)' | tail -1

Replace me.jsonl with your own agent_id.

Isolation

  • Each agent can read only its own log file. ReBAC denies cross-agent reads.
  • The mount is read-only for agents.
  • Operators with is_admin can read any agent's log.

Storage and retention

  • Backed by RAM. Default cap 10 MB per agent per day, configurable via NEXUS_ACTIVITY_AGENT_LOG_CAP_BYTES.
  • When the cap is hit, oldest lines are evicted (ring buffer). The most recent activity is always available.
  • Retention defaults to 7 days in RAM, configurable via NEXUS_ACTIVITY_AGENT_LOG_RETENTION_DAYS. No disk archive in v1.
  • NEXUS_ACTIVITY_AGENT_LOG_ENABLED=0 disables the feature entirely.
  • NEXUS_ACTIVITY_AGENT_LOG_CMD_MAX_BYTES (default 4096) controls cmd truncation length.

Metrics

  • nexus_activity_agent_log_lines_dropped_total{reason} — reason ∈ {ring_evict, recursion, no_agent}.
  • nexus_activity_agent_log_bytes{agent_id} — current per-agent buffer size.

Read the full file on GitHub · 80 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 80 lines · 0 tokens per session scan A ffabfdf6b81d

Subscribe to this mod's changes

self-observability is an agent published in the GitHub repository nexi-lab/nexus (208 stars, last pushed 3d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 743 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other agents, from other repositories

agent-architect

Use this agent when the user needs help designing, architecting, or planning new agents or workflows. This includes: Context: User wants to create a new agent system for their project. user: "I need to build an agent that monitors GitHub PRs and sends Slack notifications" assistant: "Let me use the agent-architect to…

Aurite-ai/kahuna · 300 tokens

agent-implementer

Use this agent when you need to implement AI agents based on approved implementation plans. This includes: Context: User has an approved implementation plan ready. user: "I have a plan for a GitHub PR monitoring agent, can you implement it?" assistant: "Let me use the agent-implementer to execute the implementation…

Aurite-ai/kahuna · 296 tokens

gsd-phase-researcher

Researches how to implement a phase before planning. Produces RESEARCH.md consumed by gsd-planner. Spawned by /gsd:plan-phase orchestrator.

open-gsd/gsd-core · 41 tokens

cdo

APM Chief Documentation Officer. Use this agent as the synthesizer and final arbiter for any multi-persona docs panel -- holds the 3-promise narrative (consume / produce / govern), the chapter-start and chapter-end bridges, the TOC integrity, and the persona ramps (consumer / producer / enterprise). Activate to…

microsoft/apm · 95 tokens

kg-assistant

General-purpose KG-aware assistant for any Semantica task. Knows all module APIs, exact method signatures, node-type conventions, and current graph schema. Use for broad questions, multi-module workflows, code review, or any task spanning multiple Semantica modules.

semantica-agi/semantica · 57 tokens

issue-tracker

Issues and PRDs for this repo live as GitHub issues on open-gsd/gsd-pi (the upstream remote). Use the gh CLI for all operations.

open-gsd/gsd-pi · 0 tokens