Skip to main content

process-observation

Use when emitting, querying, or rotating process-level friction observations from any skill, agent, gate, or external tool. Provides the `claude-observe` CLI one-liner, Python API `claude_observe()`, and query/sweep/compaction primitives over a two-file ledger (`active.yaml` aggregate + `events.jsonl` truth log). BEST-EFFORT writes - never raise to the caller. 13-category closed-set taxonomy. Ships with 14-day active retention, indefinite compressed stale retention, 180-day hashed event summaries, and write-time anonymization for the cross-project global rollup. S039 efficacy-telemetry extension - a gate-run denominator writer (`gate-runs.jsonl` sidecar) plus a read-only `query.py rollup` op that computes the four orchestration-efficacy metrics (gate-fail / false-positive / dual-verdict-disagreement / user-override) for alf sweeps.

Source facts

Repository
joogy06/agent-foundry
Last source activity
July 31, 2026 at 11:15
Detected SKILL.md language
English
Stars
4
Forks
0

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

File Explorer
15 files

Showing SKILL.md

SKILL.md
Source instructions · Read-only preview
nameprocess-observation
descriptionUse when emitting, querying, or rotating process-level friction observations from any skill, agent, gate, or external tool. Provides the `claude-observe` CLI one-liner, Python API `claude_observe()`, and query/sweep/compaction primitives over a two-file ledger (`active.yaml` aggregate + `events.jsonl` truth log). BEST-EFFORT writes - never raise to the caller. 13-category closed-set taxonomy. Ships with 14-day active retention, indefinite compressed stale retention, 180-day hashed event summaries, and write-time anonymization for the cross-project global rollup. S039 efficacy-telemetry extension - a gate-run denominator writer (`gate-runs.jsonl` sidecar) plus a read-only `query.py rollup` op that computes the four orchestration-efficacy metrics (gate-fail / false-positive / dual-verdict-disagreement / user-override) for alf sweeps.