inject

---
name: inject
description: Load relevant .agents context.
practices:
- wiki-knowledge-surface
- pragmatic-programmer
hexagonal_role: driving-adapter
consumes: []
produces: []
context_rel: []
skill_api_version: 1
user-invocable: true
context:
  window: fork
  intent:
    mode: none
  sections:
    exclude:
    - TASK
  intel_scope: full
metadata:
  tier: background
  dependencies:
  - compile
  - curate
  - flywheel
  internal: true
output_contract: 'stdout: injected knowledge summary; .agents/knowledge/book-of-beliefs.md, .agents/playbooks/*.md, .agents/briefings/*.md (knowledge activation)'
---
> **DEPRECATED (removal target: v3.0.0)** — Use `ao lookup --query "topic"` for on-demand learnings retrieval and phase-scoped context packets. This skill and the `ao inject` CLI command still work as compatibility adapters, but they are not the canonical context path and are not called from default hooks or other skills.

# Inject Skill

**On-demand knowledge retrieval. Not run automatically at startup (since ag-8km).**

It is read-only: it only reads knowledge for injection and never writes to `.agents/`.

Load relevant prior knowledge into the current session as a legacy adapter.

## Lease

| Field | Value |
|---|---|
| Lease | retire-candidate |
| Replacement port | `retrieve_context` / `assemble_context` |
| Replacement adapters | `ao lookup`, knowledge brief artifacts |
| Current allowed use | manual compatibility lookup only |
| Not allowed | default startup injection, hidden hook delivery, task planning |

## How It Works

In the default hookless startup path, no startup injection occurs. Run `ao session bootstrap` for the standard orientation report, then prefer `ao lookup` / `ao inject` for on-demand retrieval and bounded per-phase packets. Use `/inject` or `ao inject` only for legacy compatibility.

If you author an opt-in SessionStart hook or run a legacy hook profile, it may call:
```bash
# lean mode (MEMORY.md fresh): 400 tokens
ao inject --apply-decay --format markdown --max-tokens 400 \
  [--bead <bead-id>] [--predecessor <handoff-path>]

# legacy mode: 800 tokens
ao inject --apply-decay --format markdown --max-tokens 800 \
  [--bead <bead-id>] [--predecessor <handoff-path>]
```

This legacy path searches for relevant knowledge and prints a bounded summary.

### Work-Scoped Injection

When `--bead` is provided (via `HOOK_BEAD` env var from Gas Town):
- Learnings tagged with the same bead ID get a 1.5x score boost
- Learnings matching bead labels get a 1.2x boost
- Untagged learnings still appear but ranked lower

### Predecessor Context

When `--predecessor` is provided (path to a handoff file):
- Extracts structured context: progress, blockers, next steps
- Injected as "Predecessor Context" section before learnings
- Supports explicit handoffs, auto-handoffs, and pre-compact snapshots

## Manual Execution

Given `/inject [topic]`:

### Step 1: Search for Relevant Knowledge

**With ao CLI:**
```bash
ao lookup --query "<topic>" --limit 5
```

**Without ao CLI, search manually:**
```bash
# Global operating memory
sed -n '1,120p' ~/.agents/MEMORY.md 2>/dev/null

# Recent learnings
ls -lt .agents/learnings/ | head -5

# Recent patterns
ls -lt .agents/patterns/ | head -5

# Recent research
ls -lt .agents/research/ | head -5

# Global learnings (cross-repo knowledge)
ls -lt ~/.agents/learnings/ 2>/dev/null | head -5

# Global patterns (cross-repo patterns)
ls -lt ~/.agents/patterns/ 2>/dev/null | head -5

# Legacy patterns (read-only fallback, no new writes)
ls -lt ~/.claude/patterns/ 2>/dev/null | head -5
```

### Step 2: Read Relevant Files

Use the Read tool to load the most relevant artifacts based on topic.

### Step 3: Summarize for Context

Present the injected knowledge:
- Global principles or constraints that apply everywhere
- Key learnings relevant to current work
- Patterns that may apply
- Recent research on related topics

### Step 4: Record Citations (Feedback Loop)

After presenting injected knowledge, record which files were injected for the feedback loop:

```bash
mkdir -p .agents/ao
# Record each injected learning file as a citation
for injected_file in <list of files that were read and presented>; do
  echo "{\"artifact_path\": \"$injected_file\", \"cited_at\": \"$(date -Iseconds)\", \"session_id\": \"$(date +%Y-%m-%d)\", \"workspace_path\": \"$PWD\"}" >> .agents/ao/citations.jsonl
done
```

Citation tracking enables the feedback loop: learnings that are frequently cited get confidence boosts during `/post-mortem`, while uncited learnings decay faster.

## Knowledge Sources

| Source | Location | Priority | Weight |
|--------|----------|----------|--------|
| Global Memory | `~/.agents/MEMORY.md` | Highest | 1.0 |
| Learnings | `.agents/learnings/` | High | 1.0 |
| Patterns | `.agents/patterns/` | High | 1.0 |
| Global Learnings | `~/.agents/learnings/` | High | 0.8 (configurable) |
| Global Patterns | `~/.agents/patterns/` | High | 0.8 (configurable) |
| Research | `.agents/research/` | Medium | — |
| Retros | `.agents/learnings/` | Medium | — |
| Legacy Patterns | `~/.claude/patterns/` | Low | 0.6 (read-only, no new writes) |

## Decay Model

Knowledge relevance decays over time (~17%/week). More recent learnings are weighted higher.

## Key Rules

- **Does not run automatically** - default context delivery is explicit
- **Context-aware** - filters by current directory/topic
- **Token-budgeted** - respects max-tokens limit
- **Recency-weighted** - newer knowledge prioritized

## Examples

### Opt-In Hook Profile Invocation (legacy only)

**Hook trigger:** an externally authored or legacy `session-start.sh` may run at session start with `AGENTOPS_STARTUP_CONTEXT_MODE=lean` or `legacy`

**What happens:**
1. Hook calls `ao inject --apply-decay --format markdown --max-tokens 400` (lean) or `--max-tokens 800` (legacy)
2. CLI searches `.agents/learnings/`, `.agents/patterns/`, `.agents/research/` for relevant artifacts
3. CLI applies recency-weighted decay (~17%/week) to rank results
4. CLI outputs top-ranked knowledge as markdown within token budget
5. Agent presents injected knowledge in session context

**Result:** Prior learnings, patterns, and research are available for legacy hook profiles. This is not the default AgentOps 3.0 path.

**Note:** In the default hookless path, run `ao session bootstrap` and then pull context explicitly with `ao lookup` or `ao inject`.

### Manual Context Injection

**User says:** `/inject authentication` or "recall knowledge about auth"

**What happens:**
1. Agent calls `ao lookup --query "authentication" --limit 5`
2. CLI filters artifacts by topic relevance
3. Agent reads top-ranked learnings and patterns
4. Agent summarizes injected knowledge for current work
5. Agent references artifact paths for deeper exploration

**Result:** Topic-specific knowledge retrieved and summarized, enabling faster context loading than full artifact reads.

## Troubleshooting

| Problem | Cause | Solution |
|---------|-------|----------|
| No knowledge injected | Empty knowledge pools or ao CLI unavailable | Run `/post-mortem` to seed pools; verify ao CLI installed |
| Irrelevant knowledge | Topic mismatch or stale artifacts dominate | Use `--context "<topic>"` to filter; prune stale artifacts |
| Token budget exceeded | Too many high-relevance artifacts | Reduce `--max-tokens` or increase topic specificity |
| Decay too aggressive | Recent learnings not prioritized | Check artifact modification times; verify `--apply-decay` flag |

## Knowledge Activation (merged from `knowledge-activation`, cp-auc)

`inject` and `ao lookup` *retrieve* knowledge for the current session. **Activation** is the complementary capability — folded in here from the former `knowledge-activation` skill — that *operationalizes* a mature `.agents` corpus into durable operator surfaces (beliefs, playbooks, briefings, gaps). Where `inject` reads, activation promotes; the two are the read and write-to-surface halves of the same flywheel. Activation is the **fourth step** of the global-corpus workflow:

1. `$curate --mode=harvest` — gather artifacts from many rigs into `~/.agents/learnings/`
2. `$compile` — synthesize raw artifacts into `.agents/compiled/`
3. *(optional)* `$curate --mode=dream` overnight — bounded compounding loop
4. **knowledge activation** — lift compiled knowledge into playbooks, beliefs, and runtime briefings

`$compile` remains the hygiene loop; activation owns corpus operationalization. Use it when the problem is no longer "capture more knowledge" but: promote the strongest recurring claims into a belief system, turn healthy topics into reusable playbooks, compile a small goal-time briefing, and surface thin topics and promotion gaps before they calcify.

### Command Contract

The stable product surface is the `ao knowledge` command family:

```bash
ao knowledge activate --goal "turn agents into usable information"  # full outer loop
ao knowledge beliefs                                                # refresh belief book only
ao knowledge playbooks                                              # refresh candidate playbooks
ao knowledge brief --goal "fix auth startup"                       # goal-time briefing
ao knowledge gaps                                                   # thin topics, promotion gaps, weak claims, next work
```

`ao` owns the belief/playbook/brief/gap product surfaces directly; the skill owns routing, sequencing, interpretation, and next-step recommendations. `ao lookup` and `ao codex start` consume these outputs as operator context — matched briefings are the preferred dynamic startup surface, while selected beliefs and healthy playbooks provide bounded supporting guidance. When a retrieved briefing, belief, or playbook changes a recommendation, record it with `ao metrics cite "<path>" --type applied 2>/dev/null || true` (use `--type retrieved` for loaded-but-unused context).

### Activation Steps

1. **Preflight** — verify `.agents/` exists. To run `ao knowledge activate`, verify at least one evidence substrate is present: packet builders (`source_manifest_build.py`, `topic_packet_build.py`, `corpus_packet_promote.py`, `knowledge_chunk_build.py`) under `.agents/scripts/`; or the harvest fallback `.agents/harvest/latest.json`; or the native operator surfaces (`ao knowledge beliefs|playbooks|brief|gaps`).
2. **Consolidate evidence** — run packet layers in order: source manifests → topic packets → promoted packets → historical chunk bundles. See [references/knowledge-activation-dag.md](references/knowledge-activation-dag.md) for the full DAG and its trust gates.
3. **Distill operator surfaces** — `ao knowledge beliefs` then `ao knowledge playbooks` materialize consumer surfaces under `.agents/knowledge/` and `.agents/playbooks/`.
4. **Compile a goal-time briefing** — when there is an active objective: `ao knowledge brief --goal "..."`. Keep it small, cite source surfaces, warn when a selected topic is thin.
5. **Surface gaps** — `ao knowledge gaps` reports thin topics, missing promotions, weak claims needing review, and the next recommended mining work.
6. **Full outer loop** — `ao knowledge activate --goal "..."` sequences evidence consolidation, belief/playbook refresh, optional briefing compilation, and a gap summary in one pass.

### Activation Trust Rules

- packetization is substrate, not the product
- beliefs, playbooks, and briefings are the real operator surfaces
- thin topics stay discovery-only until evidence improves
- every generated surface should name its consumer
- repeated unchanged runs should stay structurally deterministic

### Activation Output Surfaces

Consumer-facing outputs: `.agents/knowledge/book-of-beliefs.md`, `.agents/playbooks/index.md`, `.agents/playbooks/<topic>.md`, `.agents/briefings/YYYY-MM-DD-<goal>.md`, `.agents/retro/`. Substrate surfaces: `.agents/packets/`, `.agents/topics/`, `.agents/packets/chunks/catalog.jsonl`. See [references/knowledge-activation-output-surfaces.md](references/knowledge-activation-output-surfaces.md) and [references/knowledge-activation-script-contracts.md](references/knowledge-activation-script-contracts.md) for trust boundaries and the builder inventory.

## Reference Documents

- [references/inject-cli.feature](references/inject-cli.feature) — Executable spec: `ao inject` CLI command behavior (header, JSON contract, `--for` filtering), linked to cmd tests (soc-jnfgi)
- [references/knowledge-activation.feature](references/knowledge-activation.feature) — Executable spec: consolidate evidence, distill beliefs/playbooks, compile goal-time briefing, surface gaps (soc-qk4b)
- [references/knowledge-activation-dag.md](references/knowledge-activation-dag.md) — DAG and trust gates for evidence consolidation
- [references/knowledge-activation-output-surfaces.md](references/knowledge-activation-output-surfaces.md) — canonical activation output surfaces and trust boundaries
- [references/knowledge-activation-script-contracts.md](references/knowledge-activation-script-contracts.md) — builder inventory and `ao knowledge` command ownership